MCSE Traing Guide Networking Essentials

MCSE Traing Guide Networking Essentials
01.919-x FM
8/28/98 10:08 AM
Page i
TRAINING GUIDE
MCSE
Second Edition
Networking
Essentials
Exam: 70-058
Glenn Berg
04.919-x Part1
8/28/98 10:59 AM
Page 1
P A R T
STANDARDS
AND
TERMINOLOGY
1 Networking Terms and Concepts
2 Networking Standards
I
05.919-x CH01.i
8/28/98 11:26 AM
Page 3
OBJECTIVES
Chapter 1 targets the following objectives in the
Standards and Terminology section of the Networking
Essentials exam:
Compare a client/server network with a peer-topeer network.
. This objective makes sure you are familiar with the
two main network classification models.
Define common networking terms for LANs and
WANs.
. The purpose of this objective is to make sure people working in the networking field understand the
difference between a local area network (LAN) and
a wide area network (WAN). These terms are the
main topics of discussion throughout this chapter.
Compare a file and print server with an application server.
. This objective makes sure you are aware of the different types of servers in the field of networking.
C H A P T E R
1
Networking Terms and
Concepts
05.919-x CH01.i
8/28/98 11:26 AM
Page 4
OUTLINE
Networking Concepts and Components
7
Network Services
18
9
Basic Connectivity Services
Redirector Service
Server Service
18
19
19
Centralized Computing
9
File Services
20
Distributed Computing
11
Collaborative Computing
12
File Transfer Services
Data Migration
File Archiving
File-Update Synchronization
23
24
25
25
Printing Services
26
Application Services
Database Services
Messaging/Communication Services
Email
Voice Mail
Fax Services
Groupware
26
28
30
31
31
31
31
Models of Network Computing
Network Models: Comparing
Client/Server and Peer-to-Peer
Networking Configurations
13
Client/Server Based Networking
13
Peer-to-Peer Networking
15
Local and Wide Area Networks
16
Local Area Networks (LANs)
16
Directory Services
32
Wide Area Networks (WANs)
16
Security Services
33
Intranets and Internets
17
Chapter Summary
36
05.919-x CH01.i
8/28/98 11:26 AM
Page 5
S T U DY S T R AT E G I E S
. You need to be very familiar with the terminology used throughout this chapter. This terminology serves as a basis for the rest of the book
and for the exam.
. Many different services are explained in this
book. Be prepared to understand the key
differences between a file and print server
and an application server, as well as the differences between client/server and peer-to-peer
networks. Remember that a file and print server
or an application server can be part of either a
client/server or peer-to-peer network.
. Keep in mind that this chapter presents the
big picture—a 50,000-foot overview of
networking—while at the same time introducing
basic terminology and definitions that need to
be memorized.
05.919-x CH01.i
6
8/28/98 11:26 AM
C h a p t er 1
Page 6
NE TWO R KI NG TE R M S AND CONCEPTS
INTRODUCTION
As one of the required exams in the Microsoft MCSE certification
program, the exam for Networking Essentials challenges your knowledge of computer networking components, theory, and implementation. This chapter is generic in the sense that it is not specific to any
one software or hardware vendor; instead, it introduces you to some
of the basic and rudimentary terms and concepts used when discussing networking. Real-world examples are provided whenever
possible. Study this chapter carefully; you will use these terms and
concepts throughout the rest of this book and in the real world, no
matter which networking model or system is being discussed.
Although most of this chapter’s examples are given in terms of
Microsoft solutions, all other successful networking models must
accomplish these same tasks.
This chapter begins with a definition of networking. It then moves
on to cover three different computing models used by various systems throughout the world. The discussion next turns to the two
main types of network models and then covers how networks are
classified based on various factors. The chapter goes on to describe
the various services that a network can offer.
In general, this chapter helps the reader understand some of the
broad classifications into which networks can fall. An appropriate
analogy might be motor vehicle classification—you should think in
terms of car, truck, or bus instead of a detailed description such as a
1969 Ford Mustang or a 1998 Honda Accord.
The integration of network services within personal desktop operating systems and the public emergence of the worldwide network,
also known as the Internet, have generated incredible momentum in
the movement to get connected. Networks have become the primary
means of disseminating information in most modern offices and
even in some homes.
05.919-x CH01.i
8/28/98 11:26 AM
Page 7
C hapter 1
NETWORKING CONCEPTS
COMPONENTS
NETWORKING TERMS AND CONC E PT S
7
AND
Networking is the concept of sharing resources and services. A network of computers is a group of interconnected systems sharing
resources and interacting using a shared communications link (see
Figure 1.1). A network, therefore, is a set of interconnected systems
with something to share. The shared resource can be data, a printer,
a fax modem, or a service such as a database or an email system. The
individual systems must be connected through a pathway (called the
transmission medium) that is used to transmit the resource or service
between the computers. All systems on the pathway must follow a
set of common communication rules for data to arrive at its intended destination and for the sending and receiving systems to understand each other. The rules governing computer communication are
called protocols.
In summary, all networks must have the following:
á A resource to share (resource)
á A pathway to transfer data (transmission medium)
á A set of rules governing how to communicate (protocols)
A
B
Hi, B
Hi, A
FIGURE 1.1
In its simplest form, a computer network is two
or more computers sharing information across
a common transmission medium.
05.919-x CH01.i
8
8/28/98 11:26 AM
C h a p t er 1
Page 8
NE TWO R KI NG TE R M S AND CONCEPTS
Having a transmission pathway does not always guarantee communication. When two entities communicate, they do not merely
exchange information; rather, they must understand the information
they receive from each other. The goal of computer networking,
therefore, is not simply to exchange data but to understand and use
data received from other entities on the network.
An analogy is people speaking (see Figure 1.2). Just because two people can speak, it does not mean they automatically can understand
each other. These two people might speak different languages or
interpret words differently. One person might use sign language,
while the other uses spoken language. As in human communication,
even though you have two entities who “speak,” there is no guarantee they will be able to understand each other. Just because two computers are sharing resources, it does not necessarily mean they can
communicate.
Because computers can be used in different ways and can be located
at different distances from each other, enabling computers to communicate often can be a daunting task that draws on a wide variety
of technologies.
Student
(client)
FIGURE 1.2
Human communication is like a network.
Air
(transmission
medium)
Instructor
(server)
05.919-x CH01.i
8/28/98 11:26 AM
Page 9
C hapter 1
NETWORKING TERMS AND CONC E PT S
The two main reasons for using computer networking are to provide
services and to reduce equipment costs. Networks enable computers
to share their resources by offering services to other computers and
users on a network. The following are specific reasons for networking PCs:
á Sharing files
á Sharing printers and other devices
á Enabling centralized administration and security of the
resources within the system
á Supporting network applications such as electronic mail and
database services
You will learn more about these important network functions later
in this chapter.
MODELS
OF
NETWORK COMPUTING
After you have the necessary prerequisites for network communication, a structure must be put in place that organizes how communication and sharing occurs. Three methods of organization, or
models, generally are recognized. The following are the three models
for network computing:
á Centralized computing
á Distributed computing
á Collaborative or cooperative computing
These three models are the basis for the various types of computer
networks you learn about in this book. The following sections discuss the three models for network computing.
Centralized Computing
The first computers were large, expensive, and difficult to manage.
Originally, these large mainframe computers were not networked as
you are familiar with today. Jobs were entered into the system by
reading commands from card decks. The computer executed one job
9
05.919-x CH01.i
10
8/28/98 11:26 AM
C h a p te r 1
Page 10
NE TWO R KI NG TE R MS AND CONCEPTS
at a time and generated a printout when the job was complete.
Terminals, which came later, provided the user with a new mechanism to interact with the centralized computer. These terminals,
however, were merely input/output devices that had no independent
processing power. All processing still took place on the central mainframe, (see Figure 1.3) hence the name centralized computing.
Networks, therefore, served little purpose other than to deliver commands to and get results from the powerful centralized processing
device. To this day, large mainframe systems are still being operated
around the world, most often by governments and large corporations. An example of centralized computing to which everyone can
relate is using an ATM machine. ATMs function as terminals. All
processing is done on the mainframe computer to which the ATMs
are connected. In summary, the centralized computing model
involves the following:
á All processing takes place in the central mainframe computer.
á Terminals are connected to the central computer and function
only as input/output devices.
This early computing model worked well in large organizations that
could justify the need for these expensive computing devices. One of
100% of computing
No computing
Mainframe
Dumbterminal
No computing
No computing
Printer
Dumbterminal
FIGURE 1.3
In centralized computing all the processing is
done by a central computer.
8/28/98 11:26 AM
Page 11
C hapter 1
NETWORKING TERMS AND CONCE PT S
the drawbacks, however, was that the mainframes were not flexible
in their placement (some were the size of a large room) and did not
scale down to meet the needs of smaller organizations. New ways of
sharing information were necessary to allow computing power to be
shared efficiently on smaller networks.
Distributed Computing
As personal computers (PCs) were introduced to organizations, a
new model of distributed computing emerged. Instead of concentrating computing at a central device, PCs made it possible to give each
worker an independent, individual computer. Each PC could receive
input and could process information locally, without the aid of
another computer (see Figure 1.4).
NOTE
05.919-x CH01.i
Personal Computer Terminology.
The term PC initially referred to a specific device—the IBM PC computer.
Over time, PC has become a generic
term referring to any desktop computer. Some purists, however, still use
the term PC to refer to an IBMcompatible workstation computer
and use the term Mac to refer to a
computer from Apple.
This meant that groups who previously had found the cost of a
mainframe environment to be prohibitive were now able to gain the
benefits of computing at a far lower cost than that of a mainframe.
These PCs, however, did not have the computing power of a mainframe. Thus, in most instances, a company’s mainframe could not be
replaced by a PC.
An analogy might help clarify the difference between the two computing models. A mainframe, which uses a centralized computing
model, is like a bus. A bus is a large, powerful vehicle used to transport many people at once. Everyone goes to one location—the
bus—to be transported. In the same way, everyone must work
70% of the processing
30% of the processing
Information flows
FIGURE 1.4
Distributed computing.
11
05.919-x CH01.i
12
8/28/98 11:26 AM
C h a p te r 1
Page 12
NE TWO R KI NG TE R MS AND CONCEPTS
through or at a mainframe computer. A personal PC, which uses distributed computing, is like a motorcycle. It transports one person at
a time. (Yes, I know a motorcycle can transport two people, but
think of it as only having one seat.) Each person can use his own
motorcycle to go somewhere without worrying about the other
users. PCs enable individuals to work at their own computers rather
than through a single large computer.
In summary, distributed computing involves the following:
á Multiple computers capable of processing independently
á Task completion by the local computer or other computers on
the network
Distributed computing was a major step forward in how businesses
leveraged their hardware resources. It provided smaller businesses
with their own computational capabilities, enabling them to perform
less-complex computing tasks on the smaller, relatively inexpensive
machines.
Collaborative Computing
Also called cooperative computing, collaborative computing enables
computers in a distributed computing environment to share
processing power in addition to data, resources, and services. In a
collaborative computing environment, one computer might borrow
processing power by running a program on another computer on the
network. Or, processes might be designed so they can run on two or
more computers. Collaborative computing cannot take place without a network to enable the various computers to communicate.
A person browsing the Internet is an example of collaborative computing. On the Internet, Web servers actively use resources to give
your computer information about how a Web page should look,
includings its colors, its font sizes, and what graphics should display.
Your computer uses its processing power to interpret this information and to display it in the format intended by the designer.
Another example of collaborative computing is Microsoft serverbased products such as Exchange Server or SQL Server. For both of
these products, requests originate from intelligent client software
(which uses the processor power of the workstation it is running on)
05.919-x CH01.i
8/28/98 11:26 AM
Page 13
C hapter 1
NETWORKING TERMS AND CONCE PT S
but then are serviced from server software running on a Windows
NT server. The server then processes the request using its own
resources and passes the results back to the client. Processor and
memory resources on both the client and the server are utilized in
the completion of the task.
In the future, you can expect collaborative computing to provide
even greater amounts of computing power. This might happen
through a new capability of computers to detect which PCs are idle
on the network and to harness the CPU power or RAM of the idle
PCs for use in processing.
In summary, collaborative computing involves the following:
á Multiple computers cooperating to perform a task
á Software designed to take advantage of the collaborative envi-
ronment
NETWORK MODELS: COMPARING
CLIENT/SERVER AND PEER-TO-PEER
NETWORKING CONFIGURATIONS
Compare a client/server network with a peer-to-peer network.
Networks generally fall into one of two broad network categories:
á Client/server networks
á Peer-to-peer networks
It is important to remember that one type of networking configuration is not necessarily better than another. Each type of networking
model has its own strengths and weaknesses.
Client/Server-Based Networking
A client/server network consists of a group of user-oriented PCs
(called clients) that issue requests to a server. The client PC is responsible for issuing requests for services to be rendered. The server’s
13
05.919-x CH01.i
14
8/28/98 11:26 AM
C h a p te r 1
Page 14
NE TWO R KI NG TE R MS AND CONCEPTS
function on the network is to service these requests. Servers generally
are higher-performance systems that are optimized to provide network services to other PCs. The server machine often has a faster
CPU, more memory, and more disk space than a typical client
machine.
Some examples of client/server-based networks are Novell NetWare,
Windows NT Server, and Banyan Vines. Some common server types
include file servers, mail servers, print servers, fax servers, and application servers. In a client/server network, the server machines often
are not even set up to do the tasks that a client machine can do. (On
a Novell or Banyan server, for example, a person cannot run a
spreadsheet from the server console. Other systems, such as
Windows NT and UNIX machines, enable a person to do this even
though it is not the intended use of the system).
Eating at a restaurant is analogous to a client/server model. You, the
customer, are a client. You issue requests for meals, drinks, and
dessert. The waiter is the server. It is the waiter’s job to service those
requests.
Although this discussion should have made it clear how they differ,
people often confuse mainframe computing with a client/serverbased network. The two approaches to computing are not the same,
however. In mainframe computing, the dumb terminal does not
process any requests. It simply acts as an interface to receive input
and to display output. Only the mainframe computer can process
information. In a client/server model, the client PC can process
information, but certain services are offloaded to the server machine.
The server machine’s role is simply to process the requests made for
these services by the client. In short, a client/server-based network is
one in which certain tasks run on and utilize the resources of one
machine while others utilize another machine, each according to its
functional role.
An example of a client/server system is Microsoft Exchange Server.
Your PC is responsible for constructing and displaying email messages, to name a couple of the possible tasks. The Exchange server is
responsible for delivering outgoing email and for receiving email
intended for you.
In summary, the client/server model is a network in which the role
of the client is to issue requests and the role of the server is to service
requests.
05.919-x CH01.i
8/28/98 11:26 AM
Page 15
C hapter 1
NETWORKING TERMS AND CONCE PT S
Peer-to-Peer Networking
A peer-to-peer network consists of a group of PCs that operate as
equals. Each PC is called a peer. The peers share resources (such as
files and printers) just like in a server-based network, although no
specialized or dedicated server machines exist. In short, each PC can
act as a client or a server. No one machine is set up with a higherpowered set of devices, nor is any one PC set up simply to provide
one service (such as storing files). Small networks—usually with
fewer than 10 machines—can work well in this configuration. In
larger networks, companies usually move to a server-based network
because many clients requesting to use a shared resource can put too
much strain on one client’s PC. Examples of peer-to-peer networks
include Windows for Workgroups, Windows 95, and Windows NT
Workstation.
Many actual network environments consist of a combination of
server-based and peer-to-peer networking models. In the real world,
companies often grow from a peer-to-peer network into a
client/server-based network. The following analogy might help you
better understand the use of each type of network.
A small company of 10 employees might choose to implement a carpool strategy. Let’s say four employees get together, and each takes a
turn driving the other three employees to work. This is analogous to
a peer-to-peer network. Just like a peer-to-peer network, in which no
one PC is responsible for dedicating itself to providing a service, no
one car is dedicated to providing transportation.
As the company grows to 400 employees, it might be decided that
the number of employees justifies the purchase of a dedicated ridepool van with a dedicated driver. This is analogous to a client/server
network, in which a dedicated machine is used to provide a service.
In this example, the company has dedicated a van to providing a
ride-share service.
As you can see in this analogy, no single network model fits all situations. A car pool in a small company is an efficient and cost-effective
way to get people to work. A bus probably is not economically feasible for a small company. In a big company, however, the use of a bus
becomes feasible. Peer-to-peer networks can work well for small
workgroups. Client/server networks provide the necessary resources
for larger groups of users.
15
05.919-x CH01.i
16
8/28/98 11:26 AM
C h a p te r 1
Page 16
NE TWO R KI NG TE R MS AND CONCEPTS
LOCAL
AND
WIDE AREA NETWORKS
Define common networking terms for LANs and WANs.
Networks come in all shapes and sizes. Network administrators often
classify networks according to geographical size. Networks of similar
size have many similar characteristics, as you will learn in later chapters. The following are the most common size classifications:
á Local area networks (LANs)
á Wide area networks (WANs)
These size classifications are described in the following sections.
Local Area Networks (LANs)
A local area network (LAN) is a group of computers and network
communication devices interconnected within a geographically limited area, such as a building or a campus. LANs are characterized by
the following:
á They transfer data at high speeds (higher bandwidth).
á They exist in a limited geographical area.
á Connectivity and resources, especially the transmission media,
usually are managed by the company running the LAN.
NOTE
Wide Area Networks (WANs)
WANs Are Interconnected LANs.
This interconnection often is represented by a line going into a cloud.
This is because the company running
the network typically has only a general idea of the path that the data will
take on its journey to the other LAN
segment. All the company knows is
that the data enters the cloud on one
side and exits the other side.
A wide area network (WAN) interconnects LANs. A WAN can be
located entirely within a state or a country, or it can be interconnected around the world.
WANs are characterized by the following:
á They exist in an unlimited geographical area.
á They usually interconnect multiple LANs.
á They often transfer data at lower speeds (lower bandwidth).
á Connectivity and resources, especially the transmission media,
usually are managed by a third-party carrier such as a telephone or cable company.
05.919-x CH01.i
8/28/98 11:26 AM
Page 17
C hapter 1
NETWORKING TERMS AND CONCE PT S
17
LAN
LAN
WAN Links
FIGURE 1.5
LAN
The WAN or the link up of LAN’s is often shown
as a cloud.
WANs can be further classified into two categories: enterprise WANs
and global WANs. An enterprise WAN connects the widely separated
computer resources of a single organization. An organization with
computer operations at several distant sites can employ an enterprise
WAN to interconnect the sites. An enterprise WAN can combine
private and commercial network services, but it is dedicated to the
needs of a particular organization. A global WAN interconnects
networks of several corporations or organizations. Other terms
that describe networks include municipal area network (MAN)—a
connected network that spans the geographic boundaries of a
municipality—and campus area network (CAN)—a network that
spans a campus or a set of buildings. These terms often lead to confusion because people are not sure whether they refer to the company’s own network of computers or its connection to the outside
world.
INTRANETS
AND INTERNETS
In recent years, two new terms have been introduced: internet and
intranet. A company that has a LAN has a network of computers.
As a LAN grows, it develops into an internetwork of computers,
referred to as an internet.
In the 1990s, graphical utilities (or browsers) were developed to view
information on a server. Today, the two most popular forms of this
05.919-x CH01.i
18
8/28/98 11:26 AM
C h a p te r 1
Page 18
NE TWO R KI NG TE R MS AND CONCEPTS
utility are Microsoft’s Internet Explorer and Netscape’s Navigator.
These browsers are used to navigate the Internet (note the capital I).
This terminology initially led to much confusion in the industry
because an internet is a connection of LANs, and the Internet is the
connection of servers on various LANs that is available to various
browser utilities. To avoid this confusion, the term intranet was
coined. This term describes an internetwork of computers on a LAN
for a single organization; the term Internet describes the network of
computers you can connect to using a browser—essentially, an internetwork of LANs available to the public.
NETWORK SERVICES
Network services are the basic reason we connect computers. Services
are what a company wants to have performed or provided. Based on
the services a company wants to utilize, the company purchases a
specific program and operating system. This section describes some
of the most common services available on computer networks.
Basic Connectivity Services
The PCs in a network must have special system software that enables
them to function in a networking environment. The first network
operating systems really were add-on packages that supplied the networking software for existing operating systems such as MS-DOS or
OS/2. More recent operating systems, such as Windows 95 and
Windows NT, come with the networking components built in.
An analogy might help you differentiate fully integrated systems
from add-ons. A box can hold goods, but it is not specifically
designed to go anywhere. You can place a set of logs on the ground
to act as rollers for the box, thus providing a mechanism for transporting or moving the box. This is similar to how old network systems used to work. Newer operating systems are like trucks. A truck
is designed from the ground up with a chassis that supports a box to
move goods. The box and the mechanism for transportation (the
chassis) are integrated from the beginning; they are designed to operate with each other.
05.919-x CH01.i
8/28/98 11:26 AM
Page 19
C hapter 1
NETWORKING TERMS AND CONCE PT S
19
Client and server machines require specific software components. A
computer that is strictly a server often cannot provide any client
functionality. On a Novell server or a Banyan server, for example, a
user cannot use the server for word processing. This is not always the
case, however; Microsoft’s NT Server and UNIX servers can run
client programs.
A computer in a peer-to-peer network functions as both a client and
a server; thus, it requires both client and server software. Operating
systems such as Windows NT Workstation and Windows 95, both
of which are peer-to-peer network operating systems, include dozens
of services and utilities that facilitate networking. Some of these
components are discussed in other chapters, and some are beyond
the scope of the Networking Essentials exam. (You’ll learn about
them when you study for the Windows NT Server or Windows NT
Workstation exam.) This section introduces you to a pair of key network services—the redirector service and the server service—that are
at the core of all networking functions.
Redirector Service
A network client must have a software component called a redirector.
In a typical standalone PC, I/O requests pass along the local bus to
the local CPU. The redirector intercepts I/O requests within the
client machine and checks whether the request is directed toward a
service on another computer. If it is, the redirector directs the
request toward the appropriate network entity. The redirector
enables the client machine to send information out of the computer,
provided that a transmission pathway exists.
In some operating environments, the redirector is called the
requester. The workstation service acts as a redirector on Windows
NT systems. In the field, people often refer to a redirector as a client.
To connect a Windows 95 machine to a Windows NT machine, for
example, it often is said, “Install the Microsoft Client for Microsoft
Networks.” If you want this Windows 95 machine to connect to a
Novell server, you might say, “Install a Novell Client on the
Windows 95 machine” (see Figure 1.6).
Server Service
FIGURE 1.6
A network server machine must have a component that accepts I/O
requests from clients on the network and that fulfills those requests
The dialog box on a Windows 95 machine that
shows a redirector being installed.
05.919-x CH01.i
20
8/28/98 11:26 AM
C h a p te r 1
Page 20
NE TWO R KI NG TE R MS AND CONCEPTS
by routing the requested data back across the network to the client
machine. In Windows NT, the server service performs the role of
fulfilling client requests.
File Services
Compare a file and print server with an application server.
File services enable networked computers to share files with each
other. This capability was one of the primary reasons networking of
personal computers initially came about. File services include all network functions dealing with the storage, retrieval, or movement of
data files. File services enable users to read, write, and manage files
and data. This includes moving files between computers and archiving files and data.
This section begins by defining file services and then moves on to
other related topics such as file transfers, file storage, data migration,
file archiving, and file update synchronization.
File services are an important part of client/server and peer-to-peer
networks. Computers providing files services are referred to as file
servers (see Figure 1.7). Two types of servers exist: dedicated and
non-dedicated. Dedicated servers do nothing but fulfill requests to
network clients. These servers commonly are found in client/server
environments. Non-dedicated servers do double duty. They enable a
user to go onto the machine acting as a file server and request the
use of files from other machines; at the same time, they give files to
users who request them from other computers on the network (see
Figure 1.7). Non-dedicated file servers often are found in peer-topeer networks. An example of a non-dedicated server is a Windows
95 machine that accesses files from other computers on the network
and that provides access to its hard drive for other computers.
Dedicated file servers have the following benefits:
á Files are stored in a specific place where they can be reliably
archived.
á Central file servers can be managed more efficiently because
there is a single point of storage.
á Central file servers can contain expensive high-performance
hardware that expedites file services and makes file servers
more reliable.
05.919-x CH01.i
8/28/98 11:26 AM
Page 21
C hapter 1
NETWORKING TERMS AND CONCE PT S
21
á The cost of specialized file server technology is shared by a
large number of users.
á Centralized networks are more scalable.
The following drawbacks, however, should be considered with regard
to centralized file services:
á When all data is stored on a single server, a single point of fail-
ure exists. If the server fails, all data becomes unavailable.
á Because all clients contend for file services from a single
source, average file-access times might be slower with a centralized file server than when files are stored on individual local
hard drives.
Centralized file services generally are best for organizations that want
to achieve the highest levels of centralized control for their data.
C:
DOC A
DOC B
Files
DOC C
$Idedicated servers;disadvantages>
File Server
Doc
A
Doc
B
Doc
C
FIGURE 1.7
A file server stores files for users on other network machines.
05.919-x CH01.i
22
8/28/98 11:26 AM
C h a p te r 1
Page 22
NE TWO R KI NG TE R MS AND CONCEPTS
Do not confuse centralized file services with centralized computer
models. The terms centralized and distributed in this context describe
the utilization method of processor resources, file resources, or
administrative tasks. A single administrator, for example, can watch
over a network with a single file server and many PC clients. This
network utilizes centralized administration and provides centralized
file access. Because the clients do their own processing, the network
itself fits under the distributed computing model.
In a peer-to-peer network environment, most computers can share
their files and applications with other computers, provided that a
service is installed on the machine allowing them to do this. The
sharing of services must be established for each individual computer,
and each user must have the skills required to manage the networking services on her PC. Because services are being provided by many
different computers, users must be aware of which computers are
providing which services. Clearly, the skills and responsibility
required in this situation are greater than for centralized file services.
This is in contrast to a client/server model, in which the network
often has one or more dedicated people to manage the servers.
The following are advantages of distributed file storage:
á No single point of failure exists. When a computer fails, only
the files stored on that computer become unavailable.
á Individuals typically experience faster access to files located on
their local machines than to files on centralized file servers.
á No specialized server hardware is required. File services can be
provided with standard PCs.
The following are disadvantages related to distributed file storage:
á It is more difficult to manage the file service because there is
not a single file location.
á File services provided by peers typically are not as fast or as
flexible as file services provided by a central file server specifically designed for that purpose.
á Instead of upgrading one central file server when higher per-
formance is needed, you must upgrade each computer.
Organizations tend to choose peer-to-peer networking for two reasons. The first reason is a desire to network with their current stock
05.919-x CH01.i
8/28/98 11:26 AM
Page 23
C hapter 1
NETWORKING TERMS AND CONCE PT S
of PCs without the expense of a centralized server. Another reason is
that a peer-to-peer network is an informal networking approach that
fits the working style of many organizations. Microsoft implements
peer-to-peer networking components in Windows for Workgroups,
Windows 95, and Windows NT Workstation. All of these operating
systems are capable of sharing and accessing network resources without the aid of a centralized server. These systems are not optimized
for file and printer sharing, however; this sort of network structure is
recommended only for smaller networks with limited security concerns.
File Transfer Services
Without a network, the options are limited for transferring data
between computers. You can, of course, exchange files on floppy
disks. This process is called sneaker-net because it consists of networking by physically running around and hand-delivering floppy
disks from desk to desk. Otherwise, you can use communication
software to dial up another computer and transfer files using a
modem or a direct serial connection. With a network, users have
constant access to high-speed data transfer without leaving their
desks or dialing another computer. Making a file accessible on a network is as easy as moving it into a shared directory.
Another important file-management task of the network operating
system (NOS) is providing and regulating access to programs and
data stored on the file server’s hard drive. This is known as file sharing. File sharing is another main reason companies invest in a network. Companies can save money by purchasing a single network
version of an application rather than many single-user versions.
Placing data files created by employees on a file server also serves several purposes including security, document control, and backup.
Centralized document control can be critical for a company in
which a document might need to be revised several times. In an
architectural firm, for example, the design of a building might be
created by using a drafting program such as AutoCAD. The architects might produce several versions of the building plan as the client
comes to a decision. If the plan is stored on the individual computers of each architect, the firm might not know which is the most
recent version of the plan. An older version might have the most
recent date (because of a backup, for example). If the plan is saved
23
05.919-x CH01.i
24
8/28/98 11:26 AM
C h a p te r 1
Page 24
NE TWO R KI NG TE R MS AND CONCEPTS
on a file server, however, each architect can access and work on the
same file.
Most networks have some form of centralized file storage. For many
years, companies have used the online storage approach to file storage.
In the online storage scenario, data is stored on hard disks that are
accessible on demand. The files that can be accessed on a server are
limited to the amount of available hard drive space. Hard drives are
fast, but even with drive prices decreasing in recent years, the cost to
store megabytes of data this way can still be fairly high. Hard drives
also have another disadvantage. Generally, they cannot be removed
for off-site storage or exchange or to build a library of files that are
seldom required but must be fairly readily available.
Another common approach to file storage is offline storage, which
consists of removable media that are managed manually. After data is
written to a tape or an optical disk, the storage medium can be
removed from the server and can be shelved. Users who require
offline data might need to know which tape or optical disk to
request. Some systems provide indexes or other aids that make
requesting the proper offline storage element automatic. A system
operator still has to retrieve the tape or disk, however, and mount it
on the server.
When the slow response of offline storage is unacceptable, a nearline storage approach can be used. Near-line storage employs a
machine, often called a jukebox, to manage large numbers of tapes or
optical disks automatically. The proper tape or disk is retrieved and
mounted by the jukebox without human intervention. With nearline storage, huge amounts of data can be made available with only
slight delays and at a much lower cost than storing the data on hard
drives.
Data Migration
Data migration is a technology that automatically moves infrequently
used data from online storage to near-line or offline storage. The criteria for moving files can include when the files were last used, the
owner of the files, the files’ sizes, and a variety of other factors. An
efficient data-migration facility makes it easier to locate migrated
files. Figure 1.8 illustrates one approach to data migration. Data
migration is used when dealing with near-line storage systems.
05.919-x CH01.i
8/28/98 11:26 AM
Page 25
C hapter 1
NETWORKING TERMS AND CONCE PT S
FIGURE 1.8
Data migration.
Files
used within
60 days
Files
older than
60 days
Hard
Drive
Files
older than
180 days
Optical
Jukebox
Tape
Autoloader
File Archiving
File archiving (also known as backup) is offline storage primarily
geared toward creating duplicate copies of online files. These backup
copies serve as insurance against minor or major system failures. A
redundant copy is made of important system, application, and data
files.
Generally, network administrators enable file archiving from a centralized location. A single site, for example, can back up all the
servers on a network. Many current backup systems also offer the
capability to back up various client workstations, making it feasible
to archive all files on the network to a central facility. This makes
archiving possible whether the files are located on network servers or
on the clients. This archive is then stored in a safe location. A duplicate often is made and placed off the premises in case of disaster.
File-Update Synchronization
In its simplest form, file-update synchronization ensures that all users
have the most recent copy of a file. File-update synchronization services can monitor the date and time stamps on files to determine
which files were saved most recently. By tracking the users who
access the file—along with the date and time stamps—the service
can update all copies of the file with the most recent version.
In some cases, however, file-update synchronization can be considerably more involved. In a modern computing environment, it is not
always feasible for all users to access all files in real time. A salesman,
for example, might carry a notebook computer for entering orders.
Dialing the central LAN every time an order needs to be entered is
impractical; the salesman can enter orders offline (while disconnected from the network) and can store them in the laptop. That
25
05.919-x CH01.i
26
8/28/98 11:26 AM
C h a p te r 1
Page 26
NE TWO R KI NG TE R MS AND CONCEPTS
evening, he can call the central LAN, log in, and transmit all the
day’s orders at once.
During this process, files on the LAN must be updated to reflect
new data in the salesman’s portable computer. The salesman’s PC
also might need to be updated with order confirmations or new pricing information. The process of bringing the local and remote files
into agreement also is called file-update synchronization.
File-update synchronization becomes considerably more challenging
when additional users are sharing data files simultaneously. Complex
mechanisms must be in place to make sure users do not accidentally
overwrite each other’s data. In some cases, the system simply flags
files that have multiple conflicting updates, and a human must reconcile the differences. In Windows 95 and Windows NT 4.0, the
My Briefcase program provides this service.
Printing Services
After file services, printing is probably the second biggest incentive
for installing a LAN. The following are some of the many advantages
of network print services:
á Many users can share the same printers. This capability is espe-
cially useful with expensive devices such as color printers and
plotters.
á Printers can be located anywhere, not just next to a user’s PC.
á Queue-based network printing is more efficient than direct
printing because the workstation can begin to work again as
soon as a job is queued to the network.
á Modern printing services enable users to send facsimile (fax)
transmissions through the network to a fax server.
In this book, print services are defined as a network service that controls and manages access to printers and plotters (see Figure 1.9).
Application Services
Application services enable applications to leverage the computing
power and specialized capabilities of other computers on a network.
05.919-x CH01.i
8/28/98 11:26 AM
Page 27
C hapter 1
NETWORKING TERMS AND CONCE PT S
27
FIGURE 1.9
Print services manage access to a shared
printer, making it accessible to users at other
network machines.
Print Server
Business applications, for example, often must perform complex statistical calculations beyond the scope of most desktop PCs. Statistical
software with the required capabilities might need to run on a mainframe computer or on a minicomputer. The statistical package, however, can make its capabilities available to applications on users’ PCs
by providing an application service.
The client PC sends the calculation request to the statistics server.
When the results become available, they are returned to the client.
This way, only one computer in an organization needs to have the
expensive software license and processing power required to calculate
the statistics, but all client PCs can benefit.
Application services enable organizations to install servers that are
specialized for specific functions (see Figure 1.10). Some of the more
common application servers are database servers, messaging/communication servers, groupware servers, and directory servers.
Application servers are an effective strategy for making a network
more scalable. Additional application servers can be added as new
application needs emerge. If more power is necessary for an application, only the application server needs to be upgraded. A database
server, for example, can grow from a PC to a multiprocessor RISC
05.919-x CH01.i
28
8/28/98 11:26 AM
C h a p te r 1
Page 28
NE TWO R KI NG TE R MS AND CONCEPTS
FIGURE 1.10
An application server runs all or part of an
application on behalf of a client and then
transmits the result to the client for further
processing.
(3 42 π)
x–7
2
x e –πR + 32
Application Server
=6
system running UNIX or Windows NT without requiring many (or
even any) changes to the client PCs.
If demand for a server-based application begins to affect a server’s
performance, it’s easy to move the application to a different server or
even to dedicate a server specifically to that application. This isolates
the application, enabling it and applications on the other server to
run more efficiently. This type of scalability is one of the advantages
of a LAN architecture.
Database Services
Database servers are the most common type of application servers.
Because database services enable applications to be designed in separate client and server components, such applications frequently are
called client/server databases.
With a client/server database, the client and server applications are
designed to take advantage of the specialized capabilities of client
and database systems, as described here:
05.919-x CH01.i
8/28/98 11:26 AM
Page 29
C hapter 1
NETWORKING TERMS AND CONCE PT S
á The client application manages data input from the user, gen-
eration of screen displays, some of the reporting, and dataretrieval requests sent to the database server.
á The database server manages the database files; adds, deletes,
and modifies records in the database; queries the database and
generates the results required by the client; and transmits
results back to the client. The database server can service
requests for multiple clients at the same time.
Database services relieve clients of most of the responsibilities for
managing data. A modern database server is a sophisticated piece of
software that can perform the following functions:
á Provide database security
á Optimize the performance of database operations
á Determine optimum locations for storing data without requir-
ing clients to know where the data is located
á Service large numbers of clients by reducing the amount of
time any one client spends accessing the database
á Distribute data across multiple database servers
Microsoft SQL Server and Oracle are two examples of applications
that run at the server but are able to perform tasks requested by
clients. Because of the way these applications were designed, both
require a back-end, or server, component and a front-end, or client,
component.
Distributed databases are becoming increasingly popular. They
enable portions of databases to be stored on separate server computers, which may be in different geographic locations. This technique,
known as distributed data, looks like a single logical database to
users, but it places the data users need in the most accessible location. East coast sales data, for example, might be located on a database server in Boston; West coast sales data might be on a server in
San Diego. Special database mechanisms must be in place to keep
data synchronized in the copies of the database.
More simply, databases can be replicated. Complete copies of a database can be stored in various locations. This provides a redundancy
factor because disaster is unlikely to strike all copies at once. In addition, database replication improves application response time over
29
05.919-x CH01.i
30
8/28/98 11:26 AM
C h a p te r 1
Page 30
NE TWO R KI NG TE R MS AND CONCEPTS
low-bandwidth connections because users can access the database on
the LAN rather than over a comparatively slow WAN link.
As shown in Figure 1.11, the most popular strategies for replicating
databases are the following:
á Master-driven updates. A single master server receives all
updates and, in turn, updates all replicas.
á Locally driven updates. Any local server can receive an update
and is responsible for distributing the change to other replicas.
Messaging/Communication Services
Messaging/communication services generally transfer information from
one place to another. This communication of information can be
broken down into three subareas:
á Email
á Voice mail
á Fax services
FIGURE 1.11
Replica
Replica
Master-driven and locally driven database
replications.
Replica
Replica
Replica
Update
Up
d
Update
Replica
Update
ate
Update
Up
Master
Change
Change
d
a te
Master
05.919-x CH01.i
8/28/98 11:26 AM
Page 31
C hapter 1
NETWORKING TERMS AND CONCE PT S
Email
Email systems can service any size group from a local workgroup
to a corporation to the world. By installing email routing devices,
you can transfer mail smoothly and efficiently among several LANs.
Email also can be routed to and received from the Internet. This
enables users in dozens of countries throughout the world to
exchange electronic messages.
Early text-based email has given way to elaborate systems that support embedded sound, graphics, and even video data.
Some of the major email packages include Microsoft’s Exchange
Server, Novell’s GroupWise, and Lotus Notes.
Voice Mail
Voice mail enables you to connect your computer to a telephone system and to incorporate telephone voicemail messages with your PC.
The technical term for this is telephony. This often involves moving
your voicemail messages from the phone system to the LAN and
enabling the computer network to distribute this information to different clients.
Fax Services
Fax services enable you to send or receive faxes from your computer.
This is similar to printing in that your can “print” the document to
a fax device. Fax services, however, can take on more complicated
features including the capability to send faxes to a central fax server
and to receive faxes from the phone system to a central fax device.
That device then delivers the fax message to your PC. This all occurs
automatically.
Groupware
Groupware is a relatively recent technology that enables several
network users to communicate and to cooperate when solving a
problem through shared document management. Interactive conferencing, screen sharing, and bulletin boards are examples of groupware applications. Groupware essentially is the capability for many
users to work on one or more copies of a document together.
Examples of applications with groupware features are Microsoft
Exchange, Novell’s GroupWise, and Lotus Notes.
31
05.919-x CH01.i
32
8/28/98 11:26 AM
C h a p te r 1
Page 32
NE TWO R KI NG TE R MS AND CONCEPTS
Directory Services
Directory services, also known as the x.500 standard, provide location
information for different entities on the network. Their main function is to act as an information booth, directing resource requests on
the network to the location of the resource. When a client is requesting to use a printer or to find a server or even a specific application,
the directory service tells the client where the resource is on the network and whether the resource is available (see Figure 1.12).
This is a service that more and more networking systems are moving
towards. As networking systems have developed, they have begun to
include this feature. This is similar to a large company having an
information desk, whereas a small company probably would not.
Examples of computer systems that use directory services include
Novell NetWare 4.11, Banyan VINES, Microsoft Exchange Server,
and the soon-to-be-released Windows NT 5.0.
Down the wire at address
207.219.44.3
Server
Printer
Where is the printer?
Client
FIGURE 1.12
Directory services tells clients the location of
resources on the network.
05.919-x CH01.i
8/28/98 11:26 AM
Page 33
C hapter 1
NETWORKING TERMS AND CONCE PT S
Security Services
Another service provided by networks is security. Security is one of
the most important elements involved in a network. When users
share resources and data on a network, they should be able to control who can access the data or resource and what the user can do
with it. An example of this is a file showing the financial records of a
company. If this file is on a file server, it is important to be able to
control who has access to the file. One step further, who is able to
read and change the file also is a crucial consideration. This same
example also applies to a shared printer. You might want to specify
who can use the expensive color laser printer or, more specifically,
when a person can use this printer. As you can see, security is an
important service on a network. Network administrators spend a
great deal of time learning and setting up security.
Security services often deal with a user account database or something like the aforementioned directory services. This database of
users often contains a list of names and passwords. When a person
wants to access the network, he must log on to the network.
Logging on is similar to trying to enter an office building with a
security guard at the front door. Before you can enter the building,
you must verify who you are against a list of people who are allowed
access.
Security services often are intermingled with other services. Some
services added to a network can utilize the security services of the
system onto which they have been installed. An example of this is
Microsoft Exchange Server. This messaging product can utilize the
security services of an existing Windows NT Server. An example of a
product that does not need to utilize an existing security system is
Lotus Notes. Lotus Notes has its own independent security system.
This topic is discussed in more detail in Chapter 10, “Managing and
Securing a Microsoft Network.”
N E T WO R K TE R M S I N T H E AGE OF THE I NTERNET
Computers process information. Networked computers process
information with each other. This information can be processed
centrally (mainframe), in distributed fashion, or collaboratively (network). When referring to a network of computers, the term LAN is
continues
33
05.919-x CH01.i
34
8/28/98 11:26 AM
C h a p te r 1
Page 34
NE TWO R KI NG TE R MS AND CONCEPTS
continued
used; when describing how computers are connected over large
areas, the term WAN is used. These terms often are expressed as
“Our intranet is connected to the Internet.” This translates to “Our
corporate network is connected to the global network known as the
Internet.”
One of the main reasons to have a network of computers is so
shared services are available to many users at once. These tasks
can range from storing and retrieving files to printing documents to
running databases or email. These services can be located on dedicated machines (client/server) or can be distributed on all the
client machines (peer-to-peer). In reality, a company does not say “I
want this type of network.” It simply finds a solution for its business needs. Based on this solution, the client gets a LAN or a WAN
that runs services following either a client/server or a peer-to-peer
model of networking. This enables them to process information in
some fashion.
C A S E S T U DY : M AT C H I N G N E T W O R K T Y P E
TO
C O M PA N Y N E E D S
ESSENCE OF THE CASE
SCENARIO
The following two issues are at hand:
You have an initial meeting scheduled with two
companies that have no computers.
• What details do you need from the company for you to make an informed decision?
• Based on this detailed information, will
you recommend a peer-to-peer network
or a client/server network?
All you know about these companies going into
the meeting is that each wants to install a network to increase its productivity. They want to
know whether to install a peer-to-peer network or
a client/server network. They also want to know
what details you need before a decision can be
made. You need to decide which type of network
to install in each case.
A N A LY S I S
When analyzing a computer network, it is most
important to address the types of functions the
company performs and the size of the company.
05.919-x CH01.i
8/28/98 11:26 AM
Page 35
C hapter 1
NETWORKING TERMS AND CONCE PT S
C A S E S T U DY : M AT C H I N G N E T W O R K T Y P E
Is it a small office with one or two employees?
Is it a large corporation with many people performing information gathering? Is it a company
that simply will use the network as an access
point to go out onto the Internet? What percentage of the work force uses computers?
Based on the function and size of a company,
you can determine which services it needs. This
is because a company does not buy a network,
so to speak; it purchases a business solution. A
network model is chosen based on the business
solution instead of choosing a network model
first.
The following two companies provide examples of
this principle.
The Veterinarian Clinic
A small veterinarian clinic has just set up shop in
town. This company has three employees. There
is a front desk person who books appointments
and does billing, and there are two veterinarians.
These three people need to share simple files
that make up the case file of the pet in question.
These three people also share a small printer for
printing out client bills.
A firm such as this can easily get by with a simple peer-to-peer network. Installing Windows 95
on all three machines and connecting them
together to form a small LAN provides the three
employees with the shared resources they need
to perform their job functions.
TO
C O M PA N Y N E E D S
The Large Sales Organization
A large sales organization has a huge inventory
database that is continuously updated by all 140
sales representatives. This database is central to
the existence of the firm. This company wants to
have a fax device to which all salespeople can
fax and three printers to handle all the sales
orders.
This company more than likely should go with a
client/server model. It has a large number of
people working in the office. All these people
need access to a central database, several printers, and one fax device. Ideally, this company
should purchase dedicated servers to handle
each of the three services the company wants to
incorporate into a network environment. Because
the company is very dependent on the existence
of this database, it definitely needs some form of
security service running on the servers. Strong
security services typically are found in
client/server models.
Based on the services to be provided and the
size of the organization, you can begin the
process of conceptualizing a network. As these
examples illustrate, you should start with the
services needed and work your way out to the
network model instead of jumping right into
the network topology, the operating system, and
so on.
35
05.919-x CH01.i
36
8/28/98 11:26 AM
C h a p te r 1
Page 36
NE TWO R KI NG TE R MS AND CONCEPTS
CHAPTER SUMMARY
This chapter has introduced you to a number of terms commonly
used in computer networking. It also has addressed many of the
basic networking structures you need to understand as an administrator. In doing so, this chapter has provided a general framework
you can use when analyzing a network in terms of its general design
and the function it is trying to serve or perform.
In this chapter, the exam objective “Define common networking
terms for LANs and WANs” was addressed throughout. The
“Compare a client/server network with a peer-to-peer network”
objective was covered in the section “Network Models: Comparing
Client/Server and Peer-to-Peer Networking Configurations.” Finally,
the exam objective “Compare a file and print server with an application server” was covered in the sections “File Services” and
“Application Services.”
KEY TERMS
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
network
transmission medium
protocol
centralized computing
distributed computing
collaborative or cooperative computing
client/server
peer-to-peer
local area network (LAN)
wide area network (WAN)
campus area network (CAN)
municipal area network (MAN)
Internet
intranet
redirector service
server service
file service
file transfer
data migration
file archiving
file-update synchronization
printing services
application server
database services
message/communication services
email
voice mail
fax services
groupware
directory services
security services
05.919-x CH01.i
8/28/98 11:26 AM
Page 37
C hapter 1
NETWORKING TERMS AND CONCE PT S
37
A P P LY Y O U R L E A R N I N G
The following sections enable you to assess how well
you understood the material in this chapter. The exercises provide you with opportunities to engage in the
sorts of tasks that comprise the skill sets the objectives
reflect. The review questions both review and test you
on the major concepts discussed in the chapter. The
exam questions test your knowledge of the tasks and
concepts specified in the objectives in a fashion similar
to the Microsoft exams. Answers to the review and
exam questions follow in the answers sections.
For additional review- and exam-type questions, see the
Top Score test engine on the CD-ROM that came with
this book.
Exercises
1.1
Logging On as a Peer
Objective: To explore the distinction between logging
on locally and logging on to a domain from Windows
NT Workstation. This exercise demonstrates the use of
a security service.
Estimated time: 15 minutes
1. Boot a domain-based Windows NT Workstation
computer. Press Ctrl+Alt+Del to reach the Logon
Information dialog box.
2. The box labeled Domain should display the
name of the Windows NT domain to which the
Windows NT Workstation belongs. This option
logs you in using the domain account database
located on a domain controller. Click the down
arrow to the right of the Domain box. At least
one other option—the name of the workstation
itself—should appear in the domain list. This
option logs you in using the workstation’s local
account database. The local account database is
completely separate from the domain database,
and it only gives you access to the local computer.
If the workstation is a member of a peer-to-peer
workgroup instead of a domain, the local logon
option is the only option. In fact, if a Windows
NT workstation is a member of a workgroup, the
Domain box doesn’t even appear in the Logon
Information dialog box—you automatically log
on to the local account database.
3. Select the computer name in the Domain box.
Enter a username and a password for the local
account.
If you rarely or never use the local logon option,
you may not remember a username or a password
for a local account. If you can’t remember a local
username and password, log on to the domain
from the workstation and find a local account
using the workstation’s User Manager application
(in the Administrative Tools group). Double-click
an account name to check the properties. Reset
the password if necessary. You need to log in as
Administrator to do this.
4. After you successfully log on to the local workstation account, you operate as a peer in a peer-topeer network would operate. Your credentials will
carry you no farther than the local system. Try to
access another network computer using Network
Neighborhood. Windows NT displays a dialog
box asking for a username and a password. The
computer you are accessing validates your credentials separately.
05.919-x CH01.i
38
8/28/98 11:27 AM
C h a p te r 1
Page 38
NE TWO R KI NG TE R MS AND CONCEPTS
A P P LY Y O U R L E A R N I N G
1.2
Seeing Where a Redirector Is Installed
Objective: To see where a redirector is installed on a
Windows 95 machine.
Estimated time: 10 minutes
9. Select the Cancel button three times to close all
the dialog boxes.
1.3
Exploring the NT Workstation Service
1. Power up your Windows 95 PC.
Objective: To examine the effect of stopping Windows
NT’s redirector—the Workstation service.
2. Right-click the Network Neighborhood icon and
choose the Properties option.
Estimated time: 15 minutes
3. Select the Configuration tab.
4. In the Configuration page, click the Add button.
5. Select the Client component in the Select
Network Component Type box. After you have
done this, click the Add button.
1. Log on to a Windows NT Workstation system as
an administrator.
2. Browse a shared directory on another computer
using Network Neighborhood or the Network
Neighborhood icon in Explorer. You should see a
list of the files on the shared directory.
6. The next dialog box is the Select Network Client
dialog box. This is the dialog box you interact
with when installing a redirector on Windows 95.
On the left-hand side of the dialog box is a list of
various manufacturers that have supplied
Windows 95 with redirectors to connect to their
systems. The right-hand side of the dialog box
shows a list of the redirectors, or clients, that each
vendor has supplied.
3. From the Start menu, click Settings and choose
Control Panel. Double-click the Services icon to
start the Control Panel Services application.
7. Select Microsoft on the left-hand side of the dialog box where it says Manufacturers.
5. Now try to access the shared directory using
Network Neighborhood. Without the redirector
(the Workstation service), you are unable to
access the other computers on the network.
8. On the right-hand side of the dialog box under
the heading Network Clients, you see two clients
that Microsoft supplies. (Some machines might
see three or more.) One of these clients is Client
for Microsoft Networks, a redirector to connect
Microsoft Windows 95 machines with other
Windows 95 machines and Windows NT computers. The other client, Client for NetWare
Networks, enables a Windows 95 machine to
connect to a Novell server.
4. From the Control Panel Services application,
scroll down to the Workstation service and click
the Stop button. This stops the Workstation service on your computer. Windows NT asks
whether you also want to stop some other dependent services. Click Yes.
Review Questions
1. What are three types of computing done in networks?
2. What are two main classifications of networks?
3. List five services that networks provide.
05.919-x CH01.i
8/28/98 11:27 AM
Page 39
C hapter 1
NETWORKING TERMS AND CONCE PT S
39
A P P LY Y O U R L E A R N I N G
Exam Questions
The following questions test your knowledge of the
information in this chapter. For additional exam help,
see the Top Score software on the CD-ROM that came
with this book. You also can visit Microsoft’s
Certification site at www.microsoft.com/train_cert.
1. Your client computer isn’t able to access services
on other network PCs. The problem is with your
client computer is:
4. You need to add a server to your network that
will provide services designed to alleviate the
problems caused by slow processor speeds on
many of the older machines. What type of server
will you be adding?
A. A peer
B. An application server
C. A file and print server
D. Both A and C
A. The reflector
B. The redirector
C. The server service
D. None of the above
2. You need to add a server to your domain to compensate for the shortage of disk space on many of
the older machines. What type of computer will
you be adding?
5. You are designing a small network for a single
office. The network will have nine users, each
operating from one of nine networked PCs. The
users are all accustomed to working with computers. What type of networking model is the best
solution?
A. Server-based
B. Peer-to-peer
A. A peer
C. A combination of A and B
B. An application server
D. Any of the above
C. A file and print server
D. Both A and C
3. You have a small office of computers. Each
machine is responsible for its own security. What
type of network are you running?
6. You are designing a small network for a single
office. The network will have approximately 19
users who will roam freely among the 14 participating PCs. What type of networking model is
the best solution?
A. Client/server
A. Peer-to-peer
B. Peer-to-peer
B. Cooperative
C. A combination of A and B
C. WAN
D. Any of the above
D. None of the above
05.919-x CH01.i
40
8/28/98 11:27 AM
C h a p te r 1
Page 40
NE TWO R KI NG TE R MS AND CONCEPTS
A P P LY Y O U R L E A R N I N G
7. Which type of network is most likely confined to
a building or a campus?
A. Local area
B. Metropolitan area
C. Wide area
D. Departmental
8. Which of the following can concurrently provide
and request services?
A. Server
B. Client
A. Clients request services.
B. Application services are responsible for running Microsoft Office.
C. Application servers can be optimized to specialize in a service.
D. Multiple services can be offered by the same
server PC.
12. Which three statements are true regarding database services?
A. A database server improves data security.
C. Peer
B. All data must be located on the main database
server.
D. None of the above
C. Database performance can be optimized.
9. Which file service is responsible for creating
duplicate copies of files to protect against file
damage?
A. File transfer
D. Database services enable multiple clients to
share a database.
13. Which are the two most popular strategies for
replication databases?
B. File-update synchronization
A. Offline migration
C. File archiving
B. File-update synchronization
D. Remote file access
C. Locally driven update
10. Which two of the following are file services?
A. Archiving
B. File segmenting
D. Master server update
14. Which three are advantages of a centralized
approach to providing file services?
C. Update synchronization
A. Centralized files can be readily archived.
D. Data integrity
B. It provides the best possible performance.
C. Management is efficient.
11. Which three statements are true regarding application services?
D. The cost of high-performance, high-reliability
servers can be spread across many users.
05.919-x CH01.i
8/28/98 11:27 AM
Page 41
C hapter 1
NETWORKING TERMS AND CONCE PT S
41
A P P LY Y O U R L E A R N I N G
15. Which two are advantages of a distributed
approach to providing file services?
A. There is no central point of failure.
B. It’s less difficult to manage than a complex,
centralized server.
C. It’s easily scaled to improve performance for
all users.
D. Specialized equipment is not required.
16. You want to install some services that will utilize
your network computers more efficiently.
Required Result: The bulk of the computer processing needs to be performed by the main network server.
Optional Result 1: You want all changes to the
software to be administered centrally.
Optional Result 2: Printing needs to be done
from a central location.
Suggested Solution: You install Microsoft Office
off your central file server. You also share the
printer on the central file server and provide all
your users with access.
A. This solution obtains the required result and
both optional results.
B. This solution obtains the required result and
one of the optional results.
C. This solution obtains the required result.
D. This solution does not satisfy the required
result.
17. You have advanced network users that will work
on a new network you are to install.
Required Result: Users need to be able to share
their hard drives.
Optional Result 1: Users need to be able to access
each other’s printers.
Optional Result 2: The network model should
require very little training of any new users.
Suggested Solution: Implement a peer-to-peer
network using Windows 95.
A. This solution obtains the required result and
both optional results.
B. This solution obtains the required result and
one of the optional results.
C. This solution obtains the required result.
D. This solution does not satisfy the required
result.
Answers to Review Questions
1. Three types of computing are centralized, distributed, and collaborative (cooperative) computing.
See the section “Models of Network Computing.”
2. The two main classifications of networking are
client/server and peer-to-peer networking. See
the section “Network Models: Comparing
Client/Server and Peer-to-Peer Networking
Configurations.”
05.919-x CH01.i
42
8/28/98 11:27 AM
C h a p te r 1
Page 42
NE TWO R KI NG TE R MS AND CONCEPTS
A P P LY Y O U R L E A R N I N G
3. The following are some of the possible services
provided by a network:
• File services
• Print services
• Database services
• Messaging services
• Communication services
• Security services
• Directory services
For more information on this topic, see the section
“Network Services.”
Answers to Exam Questions
1. B. A server service runs a server machine. A
redirector is run on a client machine. See the section “Basic Connectivity Services.”
2. C. A file server is a server that provides file
services to a user. An application server runs
programs for users. A file server often is called a
file and print server because it usually provides
both file and printing services. See the sections
“File Services” and “Printing Services.”
6. A. Roaming users are usually best supported by
a client/server model. See the section
“Client/Server-Based Networking.”
7. A. A local area network (LAN) usually is
defined by a network confined to a building or a
campus. See the section “Local Area Networks
(LANs).”
8. C. A peer is a machine that both provides and
requests services. See the section “Peer-to-Peer
Networking.”
9. C. File archiving, also known as tape backup, is
responsible for this. See the section “File
Services.”
10. A, C. B is a term that could be used to describe
a function of the hard drive not the file service. D
is a function of fault tolerance (discussed later in
this book). See the section “File Services.”
11. A, C, D. When Microsoft Office is stored on a
file server, the server is performing file services.
See the section “Application Services.”
12. A, C, D. Not all data must be stored on the
main database server. See the section “Database
Services.”
13. C, D. A and B are file services. See the section
“Database Services.”
3. A. Because there is no central server, this classification is known as a peer-to-peer network. See
the section “Peer-to-Peer Networking.”
14. A, C, D. A person will get faster file access if
the files are stored locally. See the section “File
Services.”
4. B. An application server is responsible for running processor-dependent applications. See the
section “Network Services.”
15. A, D. B is incorrect because it is often more
difficult to manage. C is incorrect because it is
often harder to scale. See the section “File
Services.”
5. B. A peer-to-peer network commonly is the
solution for small networks. See the section
“Peer-to-Peer Networking.”
16. D. By installing Microsoft Office on a central
file server, you are enabling all changes to the
05.919-x CH01.i
8/28/98 11:27 AM
Page 43
C hapter 1
NETWORKING TERMS AND CONCE PT S
43
A P P LY Y O U R L E A R N I N G
software to be administrated centrally, satisfying
optional result 1. By sharing out a printer—using
your printing services—you satisfy optional result
2. The required result is not met by installing
Microsoft Office on a file server; the file server is
not processing the Microsoft Office program.
The file server is simply passing files down to the
workstations. It is the workstations that are doing
all the processing.
17. B. A peer-to-peer network enables users to
share their hard drives and printers. This satisfies
the required result and optional result 1. Peer-topeer networks, however, do require more training
for users than client/server networks. This is
because users need to be trained to manage their
shared resources.
Suggested Readings and Resources
The following is recommended reading in the area
of networking terms and concepts:
1. Tanenbaum, Andrew. Computer Networks.
Prentice Hall, 1996.
2. Advances in Local and Metropolitan Area
Networks, William Stallings (editor). IEEE
Computer Society, 1994.
3. Derfler, Frank and Les Freed. How Networks
Work. Ziff Davis, 1996.
4. Hayes, Frank. Lan Times Guide to
Interoperability (Lan Times). Osborne-McGraw
Hill, 1994.
5. Wheeler, Tom, Alan Simon, and Thomas
Wheeler. Open Systems Handbook. AP
Proffessional, 1994.
05.919-x CH01.i
8/28/98 11:27 AM
Page 44
06.919-x CH02.i
8/28/98 11:46 AM
Page 45
OBJECTIVES
Chapter 2 targets the following objectives in the
Standards and Terminology section of the Networking
Essentials exam:
Define the communication devices that communicate at each level of the OSI model.
. The purpose of this exam objective is to make sure
that you are able to identify what devices on a network work within what levels of the OSI model.
Compare the implications of using connectionoriented communications with connectionless
communications.
. This exam objective addresses whether a person
understands how a connection-oriented type of
communication differs from a connectionless form
of communication.
Distinguish whether SLIP or PPP is used as the
communications protocol for various situations.
. The purpose of this objective is to make sure you
understand where, when, and for what reasons one
would use SLIP or PPP as a communications protocol.
Describe the characteristics and purpose of the
media used in IEEE 802.3 and IEEE 802.5.
. This question is asked to assess whether a person is
aware of two of the more popular implementations
of the IEEE 802.x set of standards.
C H A P T E R
2
Explain the purpose of the NDIS and Novell ODI
network standards.
. This objective is included to make sure that a person is aware of the differences between the NDIS
standard used by Microsoft networks and the ODI
standard used by Novell Networks.
Networking Standards
06.919-x CH02.i
8/28/98 11:46 AM
Page 46
OUTLINE
Standards
49
Standards Organizations and the ISO
50
Rules and the Communication Process
50
The OSI Reference Model
75
OSI Application Layer Concepts
78
Putting the OSI Model in Perspective
79
Standards That Utilize Multiple Levels
of the OSI Model
80
Serial Line Internet Protocol (SLIP) and
Point-to-Point Protocol (PPP)
80
51
How Peer OSI Layers Communicate
53
Protocol Stacks
54
Conceptualizing the Layers of the
OSI Model
OSI Presentation Layer Concepts
55
The IEEE 802 Family
OSI Physical Layer Concepts
Components That Operate at This
Level—Repeaters
OSI Data Link Layer Concepts
Hardware Access at the Data Link
Layer
Addressing at the Data Link Layer
Error and Flow Control at the Data
Link Layer
Components That Operate at This
Level—Bridges
82
56
IEEE 802.1
83
56
IEEE 802.2
83
57
IEEE 802.3
84
IEEE 802.4
84
IEEE 802.5
85
IEEE 802.6
85
IEEE 802.7
85
IEEE 802.8
85
IEEE 802.9
85
IEEE 802.10
86
IEEE 802.11
86
IEEE 802.12
86
IEEE 802.14
86
IEEE 802.3 and IEEE 802.5 Media
86
58
58
59
60
OSI Network Layer Concepts
Network Layer Addressing
Delivering Packets
Connection-Oriented and Connectionless Modes
Gateway Services
Components That Operate at This
Level—Routers
62
62
64
71
OSI Transport Layer Concepts
Transport Layer Connection Services
72
72
OSI Session Layer Concepts
Session Layer Session Administration
73
74
68
70
NDIS and ODI
88
Chapter Summary
91
06.919-x CH02.i
8/28/98 11:46 AM
Page 47
S T U DY S T R AT E G I E S
This chapter presents an overview of the OSI model
used in networking. Follow these approaches to studying the material in this chapter:
. You should be able to identify what networking
component, whether it is a device or a standard, operates at each level of the OSI model.
. You should understand the general functionality of each layer of the OSI model that is
presented.
If you understand the ODI model in these ways, you
should be ready to take this section of the exam.
06.919-x CH02.i
48
8/28/98 11:46 AM
C h a p te r 2
Page 48
NE TWO R KI NG S TANDARDS
INTRODUCTION
Before servers can provide services to clients, communications
between the two computers must be established. Beyond the cables
connecting the computers together, numerous processes operate
behind the scenes to keep things running smoothly. For these
processes to operate smoothly in a diverse networking environment,
the computing community has settled on several standards and specifications that define the interaction and interrelation of the various
components of network architecture. This chapter explores some of
those standards. It begins by exploring the Open Systems
Interconnection (OSI) reference model. This has become an industry
blueprint for defining the different components that are involved in
networking. This is an important model to learn, because all networking components and functionality are referenced within this
model. In fact, the remaining chapters of this book are organized
around the OSI model.
The chapter then moves from the OSI reference model to other
industry standards that often encompass several areas of the OSI
model at once. These standards include the Serial Line Internet
Protocol (SLIP), Point-to-Point Protocol (PPP), the IEEE 802 standards, Network Driver Interface Specification (NDIS), and Open
Data-Link Interface (ODI). The chapter concludes with a case study
applying exam-specific objectives in a real-world setting.
As noted in the study strategies for this chapter, the best approach to
the material covered in this chapter is to keep a “big picture” perspective in mind as you read through the OSI model. Focus on the
fact that the OSI model is a framework to explain concepts. This
chapter serves as a general framework, discussing general concepts
used throughout the rest of this book. Other chapters refer back to
this chapter to explain why different services and components function the way that they do. This chapter also provides a great framework for identifying and addressing real-world networking problems
and issues that may arise.
06.919-x CH02.i
8/28/98 11:47 AM
Page 49
C hapter 2
NETWORKING STANDARD S
STANDARDS
The network industry uses two types of standards: de facto standards
and de jure standards. To understand the concept of open systems
architecture, you must be familiar with the concepts of de facto and
de jure standards.
The second type of standards, de jure standards, are nonproprietary,
which means that no single company creates them or owns the rights
to them. De jure standards are developed with the intent of enhancing connectivity and interoperability by making specifications public
so that independent manufacturers can build to such specifications.
TCP/IP, which is discussed in more detail in Chapter 7, “Transport
Protocols,” is an example of a nonproprietary de jure standard.
Several permanent committees comprised of industry representatives
develop de jure standards. Some examples of these committees are
the IEEE (Institute of Electrical and Electronic Engineers) and the
IRTF (Internet Engineering Task Force). Although these committees
are supported by manufacturer subscriptions, and in some cases government representatives, they are intended to represent the interests
of the entire community and thus remain independent of any one
manufacturer’s interests. Subscribing to de jure standards reduces the
risk and cost of developing hardware and software for manufacturers.
After a standard has been finalized, a component manufacturer subscribing to it can develop products with some confidence that the
products will operate with components from other companies that
also subscribe to the same standards.
NOTE
De facto standards arise through widespread commercial and educational use. These standards often are proprietary and usually remain
unpublished and unavailable to outside vendors. Unpublished and
unavailable standards are known as closed system standards. Published
and accessible standards, on the other hand, are known as open
system standards. Through the growing acceptance of the concept
of interoperability, many closed, proprietary systems (such as IBM’s
Systems Network Architecture) have started to migrate toward
open system standards. Certainly, de facto standards are not always
closed system standards. Some examples of proprietary open system
standards include Novell’s NetWare network operating system and
Microsoft’s Windows.
Open System Standard By saying
that Microsoft has an open system
standard, this does not mean that
Microsoft has published its source
code for its products. What this
means is that Microsoft supplies
other developers with the commands
they need to enable their products to
interact with Microsoft products.
These sets of commands are also
known as Software Development Kits
or SDKs. Microsoft supplies SDKs for
virtually all its products.
49
06.919-x CH02.i
50
8/28/98 11:47 AM
C h a p te r 2
Page 50
NE TWO R KI NG S TANDARDS
An example of a de jure standard is the set of rules that guide how
web pages are transferred between computers or how files are transferred between systems. These de jure standards are created by the
IRTF to facilitate communication between different systems. One
problem of de jure standards, though, is the possibility of a vendor
choosing to follow only part of a given standard. The frequent result
is a product that claims to conform to the standard, but that in reality fails to operate with other products in the way one might believe
or expect.
Standards Organizations and the ISO
The development and implementation of de jure standards is regulated by standards organizations. For example, the CCITT (this is a
French acronym that translates to the International Consultative
Committee for Telegraphy and Telephony) and the Institute of
Electrical and Electronic Engineers (IEEE), among other organizations, are responsible for several prominent network standards that
support the International Standards Organization’s objective of network interoperability.
The International Standards Organization (ISO)—whose name is
derived from the Greek prefix iso, meaning “same”—is located in
Geneva, Switzerland. ISO develops and publishes standards and
coordinates the activities of all national standardization bodies. In
1977, the ISO initiated efforts to design a communication standard
based on the open systems architecture theory from which computer
networks would be designed. This model came to be known as the
Open Systems Interconnection (OSI) model. This model has
become an accepted framework for analyzing and developing networking components and functionality.
Rules and the Communication Process
Networks rely on many rules to manage information interchange.
Some of the procedures governed by network standards are as follows:
á Procedures used to communicate the establishment and ending
of communication
06.919-x CH02.i
8/28/98 11:47 AM
Page 51
C hapter 2
á Signals used to represent data on the transmission media
á Types of signals to be used
á Access methods for relaying a signal across the media
á Methods used to direct a message to the intended destination
á Procedures used to control the rate of data flow
á Methods used to enable different computer types to communi-
cate
á Ways to ensure that messages are received correctly
Network communication is very similar to human communication.
People follow sets of rules when they talk to one another. As a society, people have mechanisms in place to get the attention of others,
to let them know that someone is talking to them, and to establish
when they finish talking. They also have methods for verifying that
the information passed along to a person was received and understood by that person.
Like human communicaton, computer communication is an
extremely complex process, one that is often too complex to solve all
at once using just one set of rules. As a result, the industry has chosen to solve different parts of the problem with compatible standards
so that the solutions can be put together like pieces of a puzzle—a
puzzle that comes together differently each time to build a complete
communication approach for any given situation.
THE OSI REFERENCE MODEL
Having a model in mind helps you understand how the pieces of the
networking puzzle fit together. The most commonly used model is
the Open Systems Interconnection (OSI) reference model. The OSI
model, first released in 1984 by the International Standards Organization (ISO), provides a useful structure for defining and describing the various processes underlying networking communications.
The OSI model is a blueprint for vendors to follow when developing
protocol implementations. The OSI model organizes communication protocols into seven levels. Each level addresses a narrow portion of the communication process. Figure 2.1 illustrates the levels of
the OSI model.
NETWORKING STANDARD S
51
06.919-x CH02.i
52
8/28/98 11:47 AM
C h a p te r 2
Page 52
NE TWO R KI NG S TANDARDS
FIGURE 2.1
The OSI model has seven layers.
Application
All
Presentation
People
Session
Seem
Transport
To
Network
Need
Data Link
Data
Physical
Processing
NOTE
Network Medium
You should learn the names and the
order of the seven OSI layers for the
Networking Essentials exam. The following two phrases help you remember the first letters of the layers:
All People Seem To Need Data
Processing (top down)
Please Do Not Throw Sausage Pizza
Away (bottom up)
Choose one, depending on whether
you are most comfortable working
from the top of the model down or
from the bottom up.
Although you will examine each level in detail later in this chapter, a
quick overview is in order. Layer 1, the Physical layer, or Hardware
layer, as some call it, consists of protocols that control communication on the network media. Essentially, this layer deals with how
data is transferred across the transmission media. At the opposite
end, Layer 7, the Application layer, interfaces the network services
with the applications in use on the computer. These services, such
as file and print services, are discussed in Chapter 1. The five layers
in between—Data Link, Network, Transport, Session, and
Presentation—perform intermediate communication tasks. In
essence the OSI model is a framework that describes how a function
from one computer is transmitted to another computer on the network.
It is important to remember that the OSI model is not a blueprint
for how to design something; that is, it does not tell you how your
network card is suppose to operate. Instead, the OSI model is a
framework in which various networking components can be placed
into context. Many networking professionals rely on the OSI model
when troubleshooting in unfamiliar situations. These professionals
may be dealing with systems not familiar to them, but by referring
to the OSI model they are able to at least narrow down the issues at
hand.
06.919-x CH02.i
8/28/98 11:47 AM
Page 53
C hapter 2
NETWORKING STANDARD S
53
How Peer OSI Layers Communicate
Communication between OSI layers is both vertical within the OSI
layers, and also horizontal between peer layers in another computer
(see Figure 2.2). This is important to understand, because it affects
how data is passed within a computer, as well as between two computers.
When information is passed within the OSI model on a computer,
each protocol layer adds its own information to the message being
sent. This information takes the form of a header added to the
beginning of the original message. The sending of a message always
goes down the OSI stack, and hence headers are added from the top
to the bottom (see Figure 2.3).
When the message is received by the destination computer, each
layer removes the header from its peer layer. Thus at each layer headers are removed (stripped ) by the receiving computer after the information in the header has been utilized. Stripped headers are removed
in the reverse order in which they were added. That is, the last header added by the sending computer, is the first one stripped off and
read by the receiving computer.
In summary, the information between the layers is passed along vertically. The information between computers is essentially horizontal,
though, because each layer in one computer talks to its respective
layer in the other computer.
UNIX
Macintosh
Application
Application
Presentation
Presentation
Session
Session
Transport
Transport
Network
Network
Data Link
Data Link
Physical
Physical
FIGURE 2.2
Each layer in the OSI model communicates with
its peer layer on the other computer’s protocol
stack.
06.919-x CH02.i
54
8/28/98 11:47 AM
C h a p te r 2
Page 54
NE TWO R KI NG S TANDARDS
Sending
OS
Hp
Ht
Hs
Hn
Hd
Receiving
OS
Original Data
Application
Original Data
Presentation
Original Data
Session
Original Data
Network
Original Data
Data Link
Original Data
Physical
Hp
Ht
Hs
Hn
Hd
Each layer, except the Physical layer, adds a
header to the frame as it travels down the OSI
layers, and removes it as it travels up the OSI
layers.
Hp
Ht
Hs
Transport
Original Data
FIGURE 2.3
Original Data
Hn
Hd
Original Data
Original Data
Original Data
Original Data
Original Data
Original Data
= Presentation Header
= Transport Header
= Session Header
= Network Header
= Data Link Header
It should probably be noted that the Physical layer does not append
a header on to the information, because this layer deals with providing a transmission route between computers. An analogy to this is
when one sends a courier package. To send a package, you place documents into an envelope (header no. 1). This envelope is addressed
(header no. 2). The courier company places its documentation on
the package (header no. 3). This package is then moved down the
road in a vehicle (the transmission pathway). At the receiving end,
the recipient strips off the courier documentation (removing header
no. 3), then strips off the package and addressing, (the removal of
headers no. 2 and no. 1), and now has the documents at hand.
Protocol Stacks
The OSI model (and other non-OSI protocol standards) break the
complex process of network communication into layers. Each layer
represents a category of related tasks. A protocol stack is an implementation of this layered protocol architecture. The protocols and
services associated with the protocol stack interact to prepare, transmit, and receive network data.
It is important to understand just what is meant by the terms “protocol” and “protocol stack.” Often when people talk about protocols,
06.919-x CH02.i
8/28/98 11:47 AM
Page 55
C hapter 2
they mention terms such as TCP/IP or IPX. This terminology can
be misleading, for although these terms refer to protocols, they are a
specific type of protocol: transport protocols. These transport protocols often do not encompass the entire mechanism for transferring
communications. Transport protocols are discussed in Chapter 7.
Two computers must run compatible protocol stacks before they can
communicate, because each layer in one computer’s protocol stack
must interact with a corresponding layer in the other computer’s
protocol stack. For example, refer to Figure 2.2. It shows the path of
a message that starts in the Transport layer. The message travels
down the protocol stack, through the network medium, and up the
protocol stack of the receiving computer. If any layer in the receiving
computer cannot understand or is not compatible with the corresponding layer of the sending computer, the message cannot be
delivered.
To place this concept into perspective, imagine two people wishing
to communicate. If one is blind and the other is deaf, there will be a
communication problem. Both people need to convey the thought
through some form of media. However, the blind person uses voice
to transmit, which requires the receiving person to use hearing, while
the deaf person uses sign language to transmit, which requires the
receiving person to use sight.
Now if you put the idea of communicating into a layered model,
one layer constitutes the idea or need to communicate, one layer is
responsible for transmitting, and one layer is responsible for receiving the information. Both these people are using a mechanism to
transmit and receive, but the mechanisms are incompatible. In
essence, these two individuals are running different protocol stacks;
they use different systems at the layers that need to mesh.
CONCEPTUALIZING
THE OSI MODEL
THE
LAYERS
OF
The following sections provide a more detailed exposition of each of
the seven layers of the OSI model.
NETWORKING STANDARD S
55
06.919-x CH02.i
56
8/28/98 11:47 AM
C h a p te r 2
Page 56
NE TWO R KI NG S TANDARDS
OSI Physical Layer Concepts
Although the OSI Physical layer does not define the media used, this
layer is concerned with all aspects of transmitting and receiving data
on the network media. By not defining the media, this layer is not
responsible for saying whether a cable should be made of silver,
copper, or gold. Specifically, the Physical layer is concerned with
transmitting and receiving bits. This layer defines several key characteristics of the Physical network, including the following:
á Physical structure of the network (physical topology)
á Mechanical and electrical specifications for using the medium
(not the medium itself )
á Bit transmission, encoding, and timing
Although the Physical layer does not define the physical medium, it
defines clear requirements that the medium must meet. These specifications differ depending on the physical medium. Ethernet for
UTP, for example, has different specifications from coaxial ethernet.
You learn more about network transmission media in Chapter 3,
“Transmission Media.” In Chapter 4, “Network Topologies and
Architectures,” you learn more about physical topologies. This chapter is intended to give you an overview of the OSI model and which
components work at each layer. The following sections examine in
detail the components that operate at each layer, presenting this
detailed information from the bottom of the OSI model upwards.
Components That Operate at This Level—
Repeaters
Define the communication devices that communicate at each level of
the OSI model.
A repeater is a network device that repeats a signal from one port
onto the other ports to which it is connected (see Figure 2.4).
Repeaters operate at the OSI Physical layer. A repeater does not filter
or interpret anything; instead, it merely repeats (regenerates) a signal,
passing all network traffic in all directions. Signals become weaker
the farther they travel down a transmission medium, so repeaters are
used to extend the distance between network stations. The term used
to describe the loss of a signal’s strength is attenuation.
06.919-x CH02.i
8/28/98 11:47 AM
Page 57
C hapter 2
NETWORKING STANDARD S
Repeater
Weak Signal
A repeater operates at the OSI Physical layer because a repeater
doesn’t require any information from the upper layers of the OSI
model to regenerate a signal. Therefore, the repeater doesn’t have to
pass the frame to upper layers where addresses and other parameters
are interpreted. A repeater merely passes along bits of data, even if a
data frame is corrupt. The primary purpose of a repeater is to enable
the network to expand beyond the distance limitations of the transmission medium. (See Chapter 3 for more details on the lengths
associated with transmission mediums.)
The advantages of repeaters are that they are fairly inexpensive and
simple. In addition, although they cannot connect networks with
dissimilar data frames (such as a Token Ring network to an Ethernet
network), some repeaters can connect segments with similar data
frame types but dissimilar cabling (such as twisted pair and coaxial
cable).
OSI Data Link Layer Concepts
As you learned in the preceding section, the OSI Physical layer is
concerned with moving messages between two machines. Network
communication, however, is considerably more involved than moving bits from one device to another. In fact, dozens of steps must be
performed to transport a message from one device to another.
Real messages consist not of single bits but of meaningful groups of
bits. The Data Link layer receives messages, called frames, from
upper layers. A primary function of the Data Link layer is to disassemble these frames into bits for transmission and then to reconstruct the frames from the bits received.
The Data Link layer has other functions as well, such as addressing,
error control, and flow control for a single link between network
devices. (The adjacent Network layer, described later in this chapter,
Strong
Signal
(Regenerated)
FIGURE 2.4
A repeater regenerates a weak signal.
57
06.919-x CH02.i
58
8/28/98 11:47 AM
C h a p te r 2
Page 58
NE TWO R KI NG S TANDARDS
handles the more complex tasks associated with addressing and delivering packets through routers and across an internetwork.)
The IEEE 802 standard (discussed in more detail in Chapter 4,
divides the Data Link layer into two sublayers:
á Media Access Control (MAC). The MAC sublayer controls the
means by which multiple devices share the same media channel for the transmission of information. This includes contention methods (see Chapter 4), or how data is transferred
from a device, such as the network card, to the transmission
medium. The MAC layer can also provide addressing information for communication between network devices. (This is
covered in more detail in the discussion of the Network layer).
á Logical Link Control (LLC). The LLC sublayer establishes and
maintains links between communicating devices.
Hardware Access at the Data Link Layer
As the preceding section mentions, the Data Link layer’s MAC sublayer provides an interface to the network adapter card. The details
necessary to facilitate access to the network through the adapter card
are thus assigned to the Data Link layer. Some of these details
include the access control method (for example, contention or token
passing, described in Chapter 4) and the network topology.
The Data Link layer also controls the transmission method (for
example, synchronous or asynchronous) used to access the transmission medium. See Chapter 6, “Connectivity Devices and Transfer
Mechanisms,” for more on synchronous and asynchronous communications.
Addressing at the Data Link Layer
The Data Link layer maintains device addresses that enable messages
to be sent to a particular device. The addresses are called physical
device addresses. Physical device addresses are unique addresses associated with the networking hardware in the computer. In most cases
(for example, Ethernet and Token Ring), the physical device address
is burned into the NIC (network interface card) at the time the card
is manufactured. Other devices, such as ARCNet, require the changing of DIP switches on the card to set a hardware address.
06.919-x CH02.i
8/28/98 11:47 AM
Page 59
C hapter 2
The standards that apply to a particular network determine the format of the address. Because the address format is associated with the
media access control method used, physical device addresses are frequently referred to as MAC addresses.
Packets on LANs are typically transmitted so that they are available
to all devices on the network segment. Each device reads each frame
far enough to determine the device address to which the frame is
addressed. If the frame’s destination address matches the device’s own
physical address, the rest of the frame is received. If the addresses do
not match, the remainder of the packet is ignored. This is the case
for all transmissions except for those sent as broadcasts. All devices
on the network receive these broadcasts.
Bridges can be used to divide large networks into several smaller
ones. Bridges use physical device addresses to determine which
frames to leave on the current network segment and which to forward to devices on other network segments. Bridges are discussed
further later in this chapter and in even more detail in Chapter 6.
Because they use physical device addresses to manage frame routing,
bridges function at the level of the Data Link layer and are Data
Link layer connectivity devices.
Error and Flow Control at the Data Link
Layer
Several of the protocol layers in the OSI model play a role in the
overall system of flow control and error control for the network.
Flow control and error control are defined as follows:
á Flow control. Flow control determines the amount of data that
can be transmitted in a given time period. Flow control prevents the transmitting device from overwhelming the receiver.
á Error control. Error control detects errors in received frames
and requests retransmission of frames.
Error control of network communications often occurs at several different layers in the OSI model. At the Data Link layer, however,
error control consists simply of confirmation that the receiving computer got all the packets the sending computer transmitted.
Compare this to the transmission of physically shipped goods. When
NETWORKING STANDARD S
59
06.919-x CH02.i
60
8/28/98 11:48 AM
C h a p te r 2
Page 60
NE TWO R KI NG S TANDARDS
a company receives a shipment of goods one of the first things it
does is see whether the correct number of boxes arrived and whether
these boxes are damaged. This is essentially the type of error control
that happens at the Data Link layer of the OSI model. But this error
control in itself does not guarantee that the information being
received by one computer is all there. Consider the model of the
shipped boxes again: Just because all boxes arrived does not mean
that the contents of all the boxes were correctly packed or that the
merchandise in the boxes will work.
The Data Link layer’s LLC sublayer provides error control and flow
control for single links between communicating devices. The
Network layer (described in the section titled “OSI Network Layer
Concepts”) expands the system of error control and flow control to
encompass complex connections that include routers, gateways, and
internetworks.
Components That Operate at This Level—
Bridges
Define the communication devices that communicate at each level of
the OSI model.
A bridge is a connectivity device that operates at the OSI Data Link
layer. The messaging parameters available at the Data Link layer
enable a bridge to pass a frame in the direction of its destination
without simultaneously forwarding it to segments for which it was
not intended. In other words, a bridge can filter network traffic. This
filtering process reduces overall traffic because the bridge segments
the network, passing frames only when they can’t be delivered on the
local segment and passing frames to only the segment for which they
are intended.
Figure 2.5 depicts a simple bridge implementation. In this process, a
bridge filters traffic by tracking and checking the Data Link layer’s
MAC sublayer addresses of incoming frames. The bridge monitors
the source addresses of incoming frames and builds an address table
that shows which nodes are on each of the segments. When a data
frame arrives, the bridge checks the frame’s destination address and
forwards the frame to the segment that contains the destination
device or node. If the destination node exists on the same segment
as the source node, the bridge stops the frame so it doesn’t pass
06.919-x CH02.i
8/28/98 11:48 AM
Page 61
C hapter 2
NETWORKING STANDARD S
61
unnecessarily to the rest of the network. If the bridge can’t find the
destination address in its address table, it forwards the frame to all
segments except the source segment.
To understand the role a bridge plays, think of a bridge as similar to
a bridge with a toll booth on a street. The toll booth operator knows
which houses are on either side of the bridge. Based on this scenario,
when a person walks down the street and approaches the toll booth,
the toll booth operator either lets this person pass or stops him. If
this person is going to a house on the other side of the bridge, the
toll booth operator allows the person to pass. If the intended house
number is not on the other side of the bridge, the person is not
allowed to pass. Remember, the walker’s position is on the same
street the whole time. This will be important for you to remember
when the Network layer in the OSI model is discussed.
In some cases, a bridge can also perform the same functions that a
repeater performs, if this feature is built into the bridge, including
expanding cabling distance and linking dissimilar cable types. In
addition, a bridge can improve performance and reduce network
traffic by splitting the network and confining traffic to smaller segments.
Routing Table
A
This Side
This Side
A
B
C
D
C
B
D
Bridge
Broadcast
A to B
A to C
Crosses
Bridge?
Broadcast
No
Yes
No
Yes
C to D
C to A
FIGURE 2.5
Bridges isolate traffic on a single network segment.
06.919-x CH02.i
62
8/28/98 11:48 AM
C h a p te r 2
Page 62
NE TWO R KI NG S TANDARDS
So far the Physical and Data Link layers have been discussed. These
layers are concerned with connecting devices together and getting
information onto the transmission media. The next layer up in the
OSI model is concerned with how data is routed to different parts of
the network. This is a function of the Network layer.
OSI Network Layer Concepts
As you learned in the preceding section, the Data Link layer deals
with communication between devices on the same network. Physical
device addresses are used to address data frames, and each device is
responsible for monitoring the network and receiving frames
addressed to that device.
The Network layer handles communication with devices on logically
separate networks that are connected to form internetworks. Because
internetworks can be large and can be constructed of different types
of networks, the Network layer utilizes routing algorithms that guide
packets from their source to their destination networks. For more
about routing and routing algorithms, see Chapter 6.
Within the Network layer, each network in the internetwork is
assigned a network address that is used to route packets. The Network
layer manages the process of addressing and delivering packets on
internetworks.
Network Layer Addressing
You have already encountered the Data Link layer’s physical device
addresses that uniquely identify each device on a network. On larger
networks, it is impractical to deliver network data solely by means of
physical addresses. (Imagine if your network adapter had to check
every packet sent from anywhere on the Internet to look for a
matching physical address.) Larger networks require a means of routing and filtering packets to reduce network traffic and minimize
transmission time. The Network layer uses logical network addresses
to route packets to specific networks on an internetwork.
Logical network addresses are assigned during configuration of the
networks. A network installer must make sure that each network
address is unique on a given internetwork. The rules for governing
06.919-x CH02.i
8/28/98 11:48 AM
Page 63
C hapter 2
how these addresses are assigned are discussed in greater detail in
Chapter 6.
The Network layer also supports service addresses. A service address
specifies a channel to a specific process on the destination PC. The
operating systems on most computers can run several processes at
once. When a packet arrives, you must determine which process on
the computer should receive the data in the packet. You do so by
assigning service addresses, which identify upper-layer processes and
protocols. These service addresses are included with the physical and
logical network addresses in the data frame. (Some protocols refer to
service addresses as sockets or ports.)
To understand the many types of addresses used in networking, take
a step back and analyze our information so far. The analogy to be
used here is that of a house on a street in a residential neighborhood.
Imagine the address of the house is 1263 Main Street, Seattle,
Washington. As far as the postal system is concerned, all this information is the “address.” In networking, the different components
that really make up the address have names. The MAC address is
similar to the house number—1263. The network address is similar
to the street name—Main Street. Further information regarding the
address—Seattle, Washington, in this case—is analogous to the logical network address.
A service address is similar to a room in a building. If you are delivering a packet to a company, often this package needs to go one step
beyond just the front door. You can think of a service address representing a room or a department within a building, such as
Apartment 404, 1263 Main St., Seattle, Washington.
Some service addresses, called well-known addresses, are universally
defined for a given type of network. These well-known addresses are
often used for services that are shared between many different vendors. An example of this would be a web service address. Many different vendors develop web servers and web browsers. For these
components to operate with one another, a well-known address is
needed.
Other service addresses are defined by the vendors of the network
service in question. This is often the case when a vendor has some
proprietary service. In that case, only the vendor supplies the means
for communicating between the various components. An example of
this could be the service address between the cash register at a
department store and the database the cash register is updating.
NETWORKING STANDARD S
63
06.919-x CH02.i
64
8/28/98 11:48 AM
C h a p te r 2
Page 64
NE TWO R KI NG S TANDARDS
Delivering Packets
Many internetworks often include redundant data paths that you
can use to route messages. Typically, a packet passes from the local
LAN segment of the source PC through a series of other LAN segments, until it reaches the LAN segment of the destination PC. The
OSI Network layer oversees the process of determining paths and
delivering packets across the internetwork.
This is similar to when you drive from your house to work. You can
probably take a variety of routes, depending upon the events on the
roadways, such as road work or traffic jams. Based on these conditions, you choose the route to take. This type of decision-making is
what is done at the network level.
Chapter 6 describes some of the routing algorithms used to determine a path. The following sections introduce some of the basic
switching techniques. Switching techniques are mechanisms for
moving data from one network segment to another. These techniques are as follows:
á Circuit switching
á Message switching
á Packet switching
Circuit Switching
Circuit switching establishes a path that remains fixed for the duration of a connection (see Figure 2.6). Much as telephone switching
equipment establishes a route between two telephones, circuitswitching networks establish a path through the internetwork when
the devices initiate a conversation. These paths tend to be reliable
and fast in performance.
Circuit switching provides devices with a dedicated path and a welldefined bandwidth, but circuit switching is not free of disadvantages.
First, establishing a connection between devices can be timeconsuming. Second, because other traffic cannot share the dedicated
media path, bandwidth might be inefficiently utilized. This can be
compared to having a telephone conversation, yet not speaking. You
are using the line, thus not allowing others to use it, but you are not
transmitting any data. Finally, circuit-switching networks must have
a surplus of bandwidth, so these types of switches tend to be expensive to construct.
06.919-x CH02.i
8/28/98 11:48 AM
Page 65
C hapter 2
e
es
3
es
sa
S4
ge
1
M
A
Circuit switching establishes a constant path
between devices, much like a telephone connection.
M
S2
S1
S6
S3
65
FIGURE 2.6
Message 2
g
sa
NETWORKING STANDARD S
B
S5
Message Switching
Message switching treats each message as an independent entity.
Each message carries address information that describes the message’s
destination, and this information is used at each switch to transfer
the message to the next switch in the route. Message switches are
programmed with information concerning other switches in the network that can be used to forward messages to their destinations.
Message switches also may be programmed with information about
the most efficient routes. Depending on network conditions, different messages may be sent through the network by different routes, as
shown in Figure 2.7.
Message switching transfers the complete message from one switch
to the next, where the message is stored before being forwarded
again. Because each message is stored before being sent on to the
next switch, this type of network frequently is called a store-andforward network. The message switches often are general-purpose
computers and must be equipped with sufficient storage (usually
hard drives, or RAM) to enable them to store messages until forwarding is possible.
Message switching is commonly used in email because some delay is
permissible in the delivery of email. Message switching uses relatively
M
S2
S4
ge
A
sa
ge
1
2
sa
es
S1
es
S6
M
B
FIGURE 2.7
S3
S5
Message switching forwards the complete message, one switch at a time.
06.919-x CH02.i
66
8/28/98 11:48 AM
C h a p te r 2
Page 66
NE TWO R KI NG S TANDARDS
low-cost devices to forward messages and can function well with
relatively slow communication channels. Other applications for
message switching include group applications such as workflow,
calendaring, and groupware.
Message switching offers the following advantages:
á Data channels are shared among communicating devices,
improving the efficiency of available bandwidth.
á Message switches can store messages until a channel becomes
available, reducing sensitivity to network congestion.
á Message priorities can be used to manage network traffic.
á Broadcast addressing uses network bandwidth more efficiently
by delivering messages to multiple destinations.
The chief disadvantage of message switching is that message switching is not suited for real-time applications, including data communication, video, and audio.
Packet Switching
In packet switching, messages are divided into smaller pieces called
packets. Each packet includes source and destination address information so that individual packets can be routed through the internetwork independently. As you can see in Figure 2.8, the packets
that make up a message can take very different routes through the
internetwork.
So far, packet switching looks considerably like message switching,
but the distinguishing characteristic is that packets are restricted to a
size that enables the switching devices to manage the packet data
entirely in memory. This eliminates the need for switching devices to
store the data temporarily on disk. Packet switching, therefore,
routes packets through the network much more rapidly and efficiently than is possible with message switching.
Several methods of packet switching exist. Two common methods of
packet switching are as follows:
á Datagram
á Virtual circuit
These two methods are discussed in the following sections.
06.919-x CH02.i
8/28/98 11:48 AM
Page 67
C hapter 2
4
Packet switching breaks a packet up into many
different pieces that are routed independently.
S4
1
1
2 4 3 1
2
4
4 3 2 1
S1
S6
3
B
2
A
2
4 3
S3
67
FIGURE 2.8
1
S2
NETWORKING STANDARD S
S5
2
4
3
Datagram Packet Switching
Datagram services treat each packet as an independent message.
Each packet is routed through the internetwork independently, and
each switch node determines which network segment should be
used for the next step in the packet’s route. This capability enables
switches to bypass busy segments and take other steps to speed
packets through the internetwork (refer to Figure 2.8).
Datagrams are frequently used on LANs. Network layer protocols
are responsible for delivering the frame to the appropriate network.
Then, because each datagram includes destination address information (in most cases this is the MAC address), devices on the local
network can recognize and receive appropriate datagrams.
Packet switching meets the need to transmit large messages with the
fairly small frame size that can be accommodated by the Physical
layer. The Network layer is responsible for fragmenting messages
from upper layers into smaller datagrams that are appropriate for the
Physical layer. The Network layer is also responsible for reconstructing messages from datagrams as they are received.
Virtual Circuit Packet Switching
Virtual circuits operate by establishing a formal connection between
two devices in communication. When devices begin a session, they
negotiate communication parameters, such as maximum message
size, communication windows, and network paths. This negotiation
establishes a virtual circuit, which is a well-defined path through the
internetwork by which the devices communicate. This virtual circuit
generally remains in effect until the devices stop communicating.
Virtual circuits are distinguished by the establishment of a logical
connection. Virtual means that the network behaves as though a
dedicated physical circuit has been established between the communicating devices. Even though no such physical circuit actually
06.919-x CH02.i
68
8/28/98 11:48 AM
C h a p te r 2
Page 68
NE TWO R KI NG S TANDARDS
exists, the network presents the appearance of a physical connection
to the devices at the ends of the circuit.
Virtual circuits are frequently employed in conjunction with
connection-oriented services, which are discussed later in this
chapter.
Packet switching offers the following advantages:
á Packet switching optimizes the use of bandwidth by enabling
many devices to route packets through the same network channels. At any given time, a switch can route packets to several
different destination devices, adjusting the routes as required
to achieve the best efficiency.
á Because entire messages are not stored at the switches prior to
forwarding, transmission delays are significantly shorter than
those encountered with message switching.
Although the switching devices do not need to be equipped with
large amounts of hard drive capacity, they might need a significant
amount of real-time memory. In addition, the switching devices
must have sufficient processing power to run the more complex
routing protocols required for packet switching. A system must be in
place by which devices can recognize when packets have been lost so
that retransmission can be requested.
Connection-Oriented and Connectionless
Modes
Compare the implications of using connection-oriented communications with connectionless communications.
The OSI Network layer determines the route a packet will take as it
passes through a series of different LANs from the source PC to the
destination PC. The complexity and versatility of Network layer
addressing gives rise to two different communication modes for passing messages across the network, both of which are recognized under
OSI:
á Connection-oriented mode. Error correction and flow control
are provided at internal nodes along the message path.
á Connectionless mode. Internal nodes along the message path do
not participate in error correction and flow control.
06.919-x CH02.i
8/28/98 11:48 AM
Page 69
C hapter 2
To understand the distinction between connection-oriented and connectionless communications, you must consider an important distinction between the OSI model’s Data Link and Network layers. In
theory, the Data Link layer facilitates the transmission of data across
a single link between two nodes. The Network layer describes the
process of routing a packet through a series of nodes to a destination
on another link on the network. An example of this latter scenario is
a message passing from a PC on one LAN segment through a series
of routers to a PC on a distant part of the network. The internal
nodes forwarding the packet also forward other packets between
other end nodes.
In connection-oriented mode, the chain of links between the source
and destination nodes forms a kind of logical pathway connection.
The nodes forwarding the data packet can track which packet is part
of which connection. This enables the internal nodes to provide flow
control as the data moves along the path. For example, if an internal
node determines that a link is malfunctioning, the node can send a
notification message backward, through the path to the source computer. Furthermore, because the internal node distinguishes among
individual, concurrent connections in which it participates, this node
can transmit (or forward) a “stop sending” message for one of its
connections without stopping all communications through the node.
Another feature of connection-oriented communication is that internal nodes provide error correction at each link in the chain.
Therefore, if a node detects an error, it asks the preceding node to
retransmit.
Connectionless mode does not provide these elaborate internal control mechanisms; instead, connectionless mode relegates all errorcorrecting and retransmitting processes to the source and destination
nodes. The end nodes acknowledge the receipt of packets and
retransmit if necessary, but internal nodes do not participate in flow
control and error correction (other than simply forwarding messages
between the end nodes).
The advantage of connectionless mode is that connectionless communications can be processed more quickly and more simply
because the internal nodes only forward data and thus don’t have to
track connections or provide retransmission or flow control.
The differences between connection-oriented and connectionless
modes of communication may be easier to understand by analogy.
NETWORKING STANDARD S
69
06.919-x CH02.i
70
8/28/98 11:48 AM
C h a p te r 2
Page 70
NE TWO R KI NG S TANDARDS
Imagine talking to someone and then having her reaffirm that she
understood what you have told her after each sentence.
Connectionless mode is like having a conversation with someone,
but the speaker just carries on and assumes that the listener understands. Connection-oriented is slower, yet more reliable.
Connectionless is faster, but has less capability to correct errors
(misunderstandings in the conversation example) as they occur.
Connectionless mode does have its share of disadvantages, however,
including the following:
á Messages sometimes get lost due to an overflowing buffer or a
failed link along the pathway.
á If a message gets lost, the sender doesn’t receive notification.
á Retransmission for error correction takes longer because a
faulty transmission can’t be corrected across an internal link.
It is important to remember that the OSI model is not a set of
rules for communication; the OSI model is a framework in which
models of communication are explained. As such, individual implementations of connectionless protocols can attenuate some of the
preceding disadvantages. It is also important to remember that connection-oriented mode, although it places much more emphasis on
monitoring errors and controlling traffic, doesn’t always work either.
Ultimately, the choice of connection-oriented or connectionless
communications mode depends on interoperability with other systems, the premium for speed, and the cost of components.
Gateway Services
Routers can handle interconnection of networks whose protocols
function in similar ways. When the rules differ sufficiently on the
two networks, however, a more powerful device is required.
A gateway is a device that can translate the different protocols used
by different networks. Gateways can be implemented starting at the
Network layer or at higher layers in the OSI model, depending on
where the protocol translation is required.
06.919-x CH02.i
8/28/98 11:48 AM
Page 71
C hapter 2
NETWORKING STANDARD S
71
Components That Operate at This Level—
Routers
Define the communication devices that communicate at each level of
the OSI model.
A router is a connectivity device that operates at the OSI Network
layer (see Figure 2.9). The information available at the Network
layer gives a router far more sophisticated packet-delivery capabilities
than a bridge provides. As with a bridge, a router constructs a routing table, but the Network layer addressing information (discussed
earlier in this chapter) enables routers to pass packets through a
chain of other routers, or even choose the best route for a packet if
several routes exist. (See Chapter 6 for more information on routers
and how they operate.)
To understand the function of routers, it might be useful to compare
them directly to a concept you should already understand at this
point, that of a bridge. A bridge separates a LAN segment without
changing the LAN address. Think of a bridge on a street used to
cross a river. Even though you cross the bridge, you are still on the
same street. A router is more like an intersection. Think of three
PC
Network Segment 3
Network Segment 2
Router
PC
These small squares
are packets going from
segment 1 to 2.
Nothing gets into 3.
FIGURE 2.9
Network Segment 1
Routers move packets onto different segments.
06.919-x CH02.i
72
8/28/98 11:48 AM
C h a p te r 2
Page 72
NE TWO R KI NG S TANDARDS
streets converging to a single intersection. No matter which path you
take from your current street, you end up on a new street. The
router’s functionality is to direct the traffic down the correct street at
the intersection.
A hybrid device called a brouter combines some characteristics of a
router and a bridge. A brouter routes routable protocols using information available at the Network layer and acts as a bridge for nonroutable protocols. A routable protocol is a protocol that can pass
through a router. TCP/IP and IPX/SPX are examples of routable
protocols. (See Chapter 7 for more information.)
OSI Transport Layer Concepts
The Transport layer, the next layer of the OSI model, can implement
procedures to ensure the reliable delivery of messages to their destination devices. The term “reliable” does not mean that errors cannot
occur; instead, it means that if errors occur, they are detected. If
errors such as lost data are detected, the Transport layer either
requests retransmission or notifies upper-layer protocols so that they
can take corrective action.
The Transport layer enables upper-layer protocols to interface with
the network but hides the complexities of network operation from
them. One of the functions of the Transport layer is to break large
messages into segments suitable for network delivery.
Transport Layer Connection Services
Some services can be performed at more than one layer of the OSI
model. In addition to the Data Link and Network layers, the
Transport layer can take on some responsibility for connection
services. The Transport layer interacts with the Network layer’s
connection-oriented and connectionless services and provides some
of the essential quality control features. Some of the Transport layer’s
activities include the following:
á Repackaging. When large messages are divided into segments
for transport, the Transport layer must repackage the segments
when they are received before reassembling the original message.
á Error control. When segments are lost during transmission or
when segments have duplicate segment IDs, the Transport
06.919-x CH02.i
8/28/98 11:48 AM
Page 73
C hapter 2
layer must initiate error recovery. The Transport layer also
detects corrupted segments by managing end-to-end error control using techniques such as checksums.
á End-to-end flow control. The Transport layer uses acknowledg-
ments to manage end-to-end flow control between two connected devices. Besides negative acknowledgments, some
Transport layer protocols can request the retransmission of the
most recent segments.
OSI Session Layer Concepts
The next OSI layer, the Session layer, manages dialogs between two
computers by establishing, managing, and terminating communications. As illustrated in Figure 2.10, dialogs can take three forms:
á Simplex dialogs. These dialogs are responsible for one-way data
transfers only. An example is a fire alarm, which sends an
alarm message to the fire station but cannot (and does not
need to) receive messages from the fire station.
á Half-duplex dialogs. These dialogs handle two-way data trans-
fers in which the data flows in only one direction at a time.
When one device completes a transmission, this device must
“turn over” the medium to the other device so that this second
device has a turn to transmit. In a similar fashion, CB radio
operators converse on the same communication channel.
When one operator finishes transmitting, he must release his
transmit key so that the other operator can send a response.
á Full-duplex dialogs. This third type of dialog permits two-way
simultaneous data transfers by providing each device with a
separate communication channel. Voice telephones are fullduplex devices, and either party to a conversation can talk at
any time. Most computer modems can operate in full-duplex
mode.
Costs rise for half- and full-duplex operation because the more complex dialog technologies are naturally more expensive. Designers of
communications systems, therefore, generally use the simplest dialog
mode that satisfies the communication requirements.
NETWORKING STANDARD S
73
06.919-x CH02.i
74
8/28/98 11:48 AM
C h a p te r 2
Page 74
NE TWO R KI NG S TANDARDS
FIGURE 2.10
Simplex, half-duplex, and full-duplex communication modes.
Smoke
Alarm
Simplex Mode
“over”
Fire
Station
“over”
or
Half-Duplex
and
Full-Duplex
Half-duplex communication can result in wasted bandwidth during
the intervals when communication is turned around. On the other
hand, using full-duplex communication generally requires a greater
bandwidth than half-duplex communication.
The Session layer also marks the data stream with checkpoints and
monitors the receipt of those checkpoints. In the event of a failure,
the sending PC can retransmit, starting with the data sent after the
last checkpoint, rather than resend the whole message.
Session Layer Session Administration
A session is a formal dialog between a service requester and a service
provider. Sessions have at least four phases:
á Connection establishment. In this phase, a service requester
requests initiation of a service. During the setup process, communication is established and rules are agreed upon.
á Data transfer. With all the rules agreed upon during setup,
each party to the dialog knows what to expect. Communication is therefore efficient, and errors are easy to detect.
06.919-x CH02.i
8/28/98 11:48 AM
Page 75
C hapter 2
á Connection release. When the session is completed, the dialog is
terminated in an orderly fashion.
á Error Correction. Error Correction is also done at the Session
layer. It checks for errors in the reassembled packets received
from the Transport layer.
The connection establishment phase establishes the parameters for
the communication session. Actually, the connection establishment
phase is comprised of several tasks, including the following:
á Specification of required services that are to be used
á User login authentication and other security procedures
á Negotiation of protocols and protocol parameters
á Notification of connection IDs
á Establishment of dialog control, as well as acknowledgment of
numbering and retransmission procedures
After the connection is established, the devices involved can initiate
a dialog (data transfer phase). As well as exchange data, these devices
exchange acknowledgments and other control data that manage the
dialog.
The Session layer can also incorporate protocols to resume dialogs
that have been interrupted. After a formal dialog has been established, devices recognize a lost connection whenever the connection
has not been formally released. Therefore, a device realizes that a
connection has been lost when the device fails to receive an expected
acknowledgment or data transmission.
Within a certain time period, two devices can reenter the Session
that was interrupted but not released. The connection release phase
is an orderly process that shuts down communication and releases
resources on the service provider.
OSI Presentation Layer Concepts
The Presentation layer deals with the syntax, or grammatical rules,
needed for communication between two computers. The
Presentation layer converts system-specific data from the Application
NETWORKING STANDARD S
75
06.919-x CH02.i
NOTE
76
8/28/98 11:48 AM
C h a p te r 2
Page 76
NE TWO R KI NG S TANDARDS
What “Presentation” Means in the
Presentation Layer The name
“Presentation layer” has caused considerable confusion in the industry
because some people mistakenly
believe that this layer presents data
to the user. However, the name has
nothing to do with displaying data.
Instead, this function is performed by
applications running above the
Application layer.
The Presentation layer is so named
because it presents a uniform data
format to the Application layer. As a
matter of fact, this layer is not commonly implemented because applications typically perform most
Presentation layer functions.
layer into a common, machine-independent format that supports a
more standardized design for lower protocol layers.
The Presentation layer also attends to other details of data formatting, such as data encryption and data compression.
On the receiving end, the Presentation layer converts the machineindependent data from the network into the format required for the
local system. This conversion could include the following:
á Data formatting. This is the organization of the data. This
topic is actually broken down into four subtopics:
• Bit-order translation. When binary numbers are transmitted
through a network, they are sent one bit at a time. The
transmitting computer can start at either end of the number. Some computers start at the most significant digit
(MSD); others start at the least significant digit (LSD).
Essentially this has to do with whether information is read
from right to left or from left to right.
• Byte-order translation. Complex values generally must be
represented with more than one byte, but different computers use different conventions to determine which byte
should be transmitted first. Intel microprocessors, for
example, start with the least significant byte and are called
little endian. Motorola microprocessors, on the other hand,
start with the most significant byte and are called big endian. Byte-order translation might be needed to reconcile
these differences when transferring data between a computer with an Intel processor and a Motorola processor.
• Character code translation. Different computers use different binary schemes for representing character sets. For
instance: ASCII, the American Standard Code for
Information Interchange, is used to represent English characters on all microcomputers and most minicomputers (see
Figure 2.11); EBCDIC, the Extended Binary Coded
Decimal Interchange Code, is used to represent English
characters on IBM mainframes (see Figure 2.12); and
Shift-JIS is used to represent Japanese characters.
• File syntax translation. File formats differ between computers. For instance, Macintosh files actually consist of two
06.919-x CH02.i
8/28/98 11:49 AM
Page 77
C hapter 2
77
NETWORKING STANDARD S
FIGURE 2.11
ASCII character code, used by PCs.
8
7
6
5
4
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
3
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
2
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
0
0
0
NUL
SOH
STX
ETX
PF
HT
LC
DEL
0
0
0
1
DLE
DC1
DC2
DC3
RES
NL
BS
IL
CAN
EM
SMM CC
VT
FF
IFS
CR IGS
SO IRS
SI
IUS
0
0
1
0
0
0
1
1
DS
SOS
FS SYN
0
1
0
0
0
1
0
1
0
1
1
0
SP
&
—
/
0
1
1
1
1
0
0
0
1
0
0
1
j
k
l
m
n
o
p
q
r
,
%
>
?
a
b
c
d
e
f
g
h
i
:
#
@
,
=
"
BYP PN
LF
RS
EOB UC
PRE EOT
SM
DC4
ENQ NAK
ACK
BEL SUB
.
<
(
+
|
!
$
*
)
;
1
0
1
0
s
t
u
v
w
x
y
z
1
0
1
1
1
1
0
0
1
1
0
1
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
1
1
1
0
1
1
1
1
S
T
U
V
W
X
Y
Z
0
1
2
3
4
5
6
7
8
9
FIGURE 2.12
EDIBIC character code, used by IBM mainframes.
06.919-x CH02.i
78
8/28/98 11:49 AM
C h a p te r 2
Page 78
NE TWO R KI NG S TANDARDS
related files called a data fork and a resource fork. PC files,
on the other hand, consist of a single file.
NOTE
á Encryption. Encryption puts data into a form unreadable by
Many vendors incorporate Unicode in
their products. Unicode, a 16-bit code
that can represent 65,536 characters
in English and other languages, is
organized into code pages devoted to
the characters required for a given
language. Unicode improves the
portability of products between different language environments.
unauthorized users. Encryption takes on two main forms:
• Public key. This uses a rule of encryption (the key) and a
known value. The manipulation of the key with a known
value produces a mechanism for decrypting data.
• Private key. This encryption uses one key. All components
that have the key can decrypt the data.
The redirector service operates at the OSI Presentation layer.
OSI Application Layer Concepts
The Application layer of the OSI reference model is concerned
with providing services on the network, including file services, print
services, application services such as database services, messaging services, and directory services among others.
A common misunderstanding is that the Application layer is responsible for running user applications such as word processors. This is
not the case. The Application layer, however, does provide an interface whereby applications can communicate with the network. It is
this interface that is often referred to as the Application
Programming Interface (API). Some examples of APIs include MAPI
(Messaging Programming Interface) and TAPI (Telephony
Application Programming Interface).
The Application layer also advertises the available services that your
computer has to the network. An example of this is when you
double-click on the Network Neighborhood Icon in Windows 95 or
Windows NT. The resulting picture shows a list of computers that
have services available to network users. (The security service of
these computers determines whether or not a user has access.)
The nature of these services is beyond the scope of this chapter. In
general, you would take a course or read a book to understand how
to manage different services on a network. Examples of these services
include Windows NT security, file and printing management,
SMTP mail, fax servers, and ODBC connectivity. This book covers
some of the basics of Windows NT Security in Chapter 10,
“Managing and Securing a Microsoft Network.”
06.919-x CH02.i
8/28/98 11:49 AM
Page 79
C hapter 2
PUTTING THE OSI MODEL
PERSPECTIVE
IN
To put the whole OSI model into context, begin by assuming that
you are sitting at your PC, using Microsoft Word, and you have just
elected to print the document to a printer attached to your neighbor’s computer.
The word processor sends the print job down to the redirector
(Presentation layer). From here it goes to get information about its
destination. This information is contained in the Session layer, where
a session was established by your neighbor’s computer. At this point
the data is broken up into smaller chunks of information (Transport
layer). From here it is addressed, so it can get to the other computer
(Network Layer). At this point it is sent to the network card (Data
Link layer) so that the print job can be converted into signals that
run down the network cable (Physical layer).
From this point it may pass between several repeaters if the network
cable is going a long distance. The packets of data may also go
through various bridges if they are on a network segment that is
heavily populated with machines. These packets containing the print
job may also go through various routers, if the destination printer is
on another network segment in the office.
At this point, the packets containing the print job arrive at the network card of the destination computer with the printer. From here
the signals from the network cable are converted back into a format
that the computer can understand (Data Link layer). Then the information on the intended address is verified as indicating the receiving
computer (Network layer). At this point packets are reassembled to
form a proper job (Transport layer). During the receiving of the
packets, the destination computer must know when it has received
the complete the print job (Session layer). Finally, the information is
presented to the Print Service (the Application layer).
In the real world, you usually can pick individual OSI layer components at the Physical (transmission media types, repeaters), Data
Link (network card, network card drivers and bridges), Network
(routers and brouters) and Application (services you wish to have on
your network) levels. Typically, the Transport, Session, and
Presentation layers are built into the networking components of
NETWORKING STANDARD S
79
06.919-x CH02.i
80
8/28/98 11:49 AM
C h a p te r 2
Page 80
NE TWO R KI NG S TANDARDS
operating systems or programs that are used and do not present a set
of alternatives from which to choose.
STANDARDS THAT UTILIZE MULTIPLE
LEVELS OF THE OSI MODEL
The discussion in most of this chapter has focused on the explanation of the OSI model and its seven levels. This chapter has discussed what occurs at each level, and in some cases, has given
examples of components that operate at these different levels.
The remainder of this chapter looks at some other standards or protocols that are common features of networks. These standards often
encompass several layers of the OSI model at once. The three broad
standards that will be examined are the following:
á SLIP and PPP
á The IEEE 802 suite of standards
á NDIS and ODI
When working with Microsoft networks, you will come across these
standards on more than one occasion when dealing with connectivity issues.
Serial Line Internet Protocol (SLIP)
and Point-to-Point Protocol (PPP)
Distinguish whether SLIP or PPP is used as the communication protocol for various situations.
Two other standards vital to network communication are Serial Line
Internet Protocol (SLIP) and Point-to-Point Protocol (PPP). SLIP and
PPP were designed to support dial-up access to networks based on
the Internet transport protocols. SLIP is a simple protocol that functions at the Physical layer, whereas PPP is a considerably enhanced
protocol that provides Physical layer and Data Link layer functionality. The relationship of both to the OSI model is shown in Figure
2.13.
06.919-x CH02.i
8/28/98 11:49 AM
Page 81
C hapter 2
NETWORKING STANDARD S
81
FIGURE 2.13
Application
The relationship between SLIP, PPP, and the OSI
model.
Presentation
Session
Transport
Network
Data Link
SLIP
PPP
Developed to provide dial-up TCP/IP connections, SLIP is an
extremely rudimentary protocol that suffers from a lack of rigid standardization in the industry, which sometimes hinders different vendor implementations of SLIP from operating with each other.
Windows NT supports both SLIP and PPP from the client end
using the Dial-Up Networking application. On the server end,
Windows NT RAS (Remote Access Service) supports PPP but
doesn’t support SLIP. In other words, Windows NT can act as a PPP
server but not as a SLIP server.
SLIP is most commonly used on older systems or for dial-up connections to the Internet via SLIP-server Internet hosts.
PPP was defined by the Internet Engineering Task Force (IETF) to
improve on SLIP by providing the following features:
á Security using password logon
á Simultaneous support for multiple protocols on the same link
á Dynamic IP addressing
á Improved error control
Different PPP implementations might offer different levels of service
and negotiate service levels when connections are made. Due to its
versatility, interoperability, and additional features, PPP is presently
surpassing SLIP as the most popular serial-line protocol.
Certain dial-up configurations cannot use SLIP for the following
reasons:
á SLIP supports the TCP/IP transport protocol only. PPP, how-
ever, supports TCP/IP, as well as a number of other transport
NOTE
Physical
RAS and Dial-Up Networking
Windows NT RAS is a dial-up service
that ships with Windows NT. This service is known as Dial-Up Networking
in Windows 95, and essentially
enables one to connect computer systems using telephone lines. Anytime
you dial up an ISP (Internet Service
Provider), you are experiencing functionality similar to RAS.
06.919-x CH02.i
82
8/28/98 11:49 AM
C h a p te r 2
Page 82
NE TWO R KI NG S TANDARDS
protocols, such as NetBEUI, IPX, AppleTalk, and DECnet. In
addition, PPP can support multiple protocols over the same
link.
á SLIP requires static IP addresses. Because SLIP requires
static—or preconfigured—IP addresses, SLIP servers do not
support Dynamic Host Configuration Protocol (DHCP),
which assigns IP addresses dynamically, or when requested.
(DHCP enables clients to share IP addresses so that a relatively
small number of IP addresses can serve a larger user base.) If
the dial-up server uses DHCP to assign an IP address to the
client, the dial-up connection won’t use SLIP.
á SLIP does not support dynamic addressing through DHCP.
SLIP connections, therefore, cannot dynamically assign a
WINS or DNS server.
Windows NT RAS (using PPP) offers a number of other interesting
features, including the following:
á PPP Multilink Protocol. Multilink enables a single connection
to use several physical pathways of the same type (such as
modems, ISDN lines, and X.25 cards). Utilizing multiple
pathways for a single connection increases bandwidth and,
therefore, performance.
á NetBIOS Gateway. A RAS server can connect a client running
the NetBEUI protocol with a TCP/IP or IPX network by serving as a NetBIOS gateway.
á IPX or IP Router. A RAS server can act as a router for IPX/SPX
and TCP/IP networks. (See Chapter 6 for more information
on routers.)
NOTE
THE IEEE 802 FAMILY
Origin of the 802 Number The IEEE
decided to name the 802 family set
of standards “802” because it was in
February of 1980 that they started
this project of standardization.
The Institute of Electrical and Electronic Engineers (IEEE) is one of
the largest professional organizations in the world, and is extremely
influential with regard to setting standards. In February of 1980, the
IEEE implemented a task force to develop a set of standards for connectivity between Network Interface Cards (NICs) and transmission
media. This task force was known as the 802 committee. This 802
committee was broken down into several different subcommittees
06.919-x CH02.i
8/28/98 11:49 AM
Page 83
C hapter 2
NETWORKING STANDARD S
that were each responsible for some different implementation of
data transfer that occurs at the Data Link level of the OSI model.
These IEEE standards have also been adopted by ISO, and they
are referred to as ISO 8802. The IEEE 802 series of standards, as
well as all the other IEEE standards and research, can be found at
http://standards.ieee.org/802/index.html.
Thirteen workgroups oversee the 802 standards. Each workgroup is
assigned a specific mandate in the area of LAN/MAN connectivity
they are to analyze. Figure 2.14 illustrates the position each standard
occupies in the OSI reference model.
IEEE 802.1
This standard is actually one that goes beyond the Data Link layer
of the OSI model. This is a general standard for network management, and provides network management standards to the other 802
standards in the OSI model. This standard actually covers all layers
from the Physical to the Transport layer.
IEEE 802.2
The IEEE 802.2 standard defines an LLC sublayer that is used by
other lower-layer protocols. Because these lower-layer protocols can
use a single LLC protocol layer, Network layer protocols can be
designed independently of both the network’s Physical layer and
MAC sublayer implementations.
Application
Presentation
Session
Transport
802.10
Network
802.2
Data Link
Physical
802.3
802.4
802.5
802.6
802.1
802.11
802.12
802.9 (proposed) (proposed)
FIGURE 2.14
The relationship between the IEEE 802 standards and the OSI model.
83
06.919-x CH02.i
84
8/28/98 11:49 AM
C h a p te r 2
Page 84
NE TWO R KI NG S TANDARDS
The LLC appends to packets a header that identifies the upper-layer
protocols associated with the frame. This header also declares the
processes that are the source and destination of each packet.
The workgroup for this set of standards is currently inactive.
IEEE 802.3
The IEEE 802.3 standard defines a network derived from the ethernet network originally developed by Digital, Intel, and Xerox. This
standard defines characteristics related to the MAC sublayer of the
Data Link layer and the OSI Physical layer. With one minor distinction—frame type—IEEE 802.3 Ethernet functions identically to
DIX Ethernet v.2. These two standards can even coexist on the same
cabling system, although devices using one standard cannot communicate directly with devices using the other.
The MAC sublayer uses a type of contention access called Carrier
Sense Multiple Access with Collision Detection (CSMA/CD). This
technique reduces the incidence of collision by having each device
listen to the network to determine whether it’s quiet (“carrier sensing”); a device attempts to transmit only when the network is quiescent. This reduces but does not eliminate collisions because signals
take some time to propagate through the network. As devices transmit, they continue to listen so they can detect a collision should it
occur. When a collision occurs, all devices cease transmitting and
send a “jamming” signal that notifies all stations of the collision.
Then, each device waits a random amount of time before attempting
to transmit again. This combination of safeguards significantly
reduces collisions on all but the busiest networks.
IEEE 802.4
The 802.4 standard describes a network with a bus physical topology that controls media access with a token mechanism. This standard
was designed to meet the needs of industrial automation systems but
has gained little popularity. Both baseband and broadband (using
75-ohm coaxial cable) configurations are available.
The workgroup for this set of standards is currently inactive.
06.919-x CH02.i
8/28/98 11:49 AM
Page 85
C hapter 2
IEEE 802.5
The IEEE 802.5 standard was derived from IBM’s Token Ring network, which employs a ring logical topology and token-based media
access control. Data rates of 1, 4, and 16Mbps have been defined for
this standard. More discussion on Token Ring will be seen in
Chapter 4.
IEEE 802.6
The IEEE 802.6 standard describes a MAN standard called
Distributed Queue Dual Bus (DQDB). Much more than a data network technology, DQDB is suited to data, voice, and video transmissions. The network is based on fiber-optic cable in a dual-bus
topology, and traffic on each bus is unidirectional. When operated in
pairs, the two buses provide a fault-tolerant configuration.
Bandwidth is allocated by using time slots, and both synchronous
and asynchronous modes are supported.
The workgroup for this set of standards is currently inactive.
IEEE 802.7
This standard deals with integrating broadband solutions into a network environment. This standard is currently under development.
The workgroup for this set of standards is currently inactive.
IEEE 802.8
This standard deals with methods of implementing fiber optic technology into networking environments. This standard is currently
under development.
IEEE 802.9
The IEEE 802.9 standard supports a 10Mbps asynchronous channel, along with 96 64Kbps (6Mbps total bandwidth) channels that
can be dedicated to specific data streams. The total bandwidth is
NETWORKING STANDARD S
85
06.919-x CH02.i
86
8/28/98 11:49 AM
C h a p te r 2
Page 86
NE TWO R KI NG S TANDARDS
16Mbps. This standard is called Isochronous Ethernet (IsoEnet) and is
designed for settings with a mix of bursty and time-critical traffic.
IEEE 802.10
This standard deals with security and encryption standards. This
standard is currently under development.
IEEE 802.11
IEEE 802.11 is a standard for wireless LANs and is currently under
development. A CSMA/CD method has been approved, but the
final standard is pending.
IEEE 802.12
The IEEE 802.12 standard is based on a 100Mbps proposal promoted by AT&T, IBM, and Hewlett-Packard. Called 100VG-AnyLAN,
the network is based on a star-wiring topology and a contentionbased access method whereby devices signal the wiring hub when
they need to transmit data. Devices can transmit only when granted
permission by the hub. This standard is intended to provide a highspeed network that can operate in mixed ethernet and token-ring
environments by supporting both frame types.
IEEE 802.14
The 802.13 designation is not used, hence the last standard is
known as 802.14. This standard is for transmitting data over cable
TV lines. The committee is currently looking at a hybrid fiber/coax
media. This is one of the up and coming areas for fast Internet
access from a person’s home.
IEEE 802.3 and IEEE 802.5 Media
Describe the characteristics and purpose of the media used in the
IEEE 802.3 and IEEE 802.5.
06.919-x CH02.i
8/28/98 11:49 AM
Page 87
C hapter 2
NETWORKING STANDARD S
IEEE 802.2 (topology independent), IEEE 802.3 (based on
Ethernet), and IEEE 802.5 (based on token ring) are the most commonly used IEEE 802 standards. Carefully read the following
overview of the media each uses; Microsoft expects you to describe
the characteristics and purpose of the media used in IEEE 802.3 and
IEEE 802.5 for the Networking Essentials exam. (Chapters 3 and 4
discuss Ethernet and token-ring media in greater detail.)
The following list details the IEEE 802.3 variants of transmission
media:
á lBASE5. This 1Mbps network utilizes UTP cable with a signal
range up to 500 meters (250 meters per segment). A star physical topology is used.
á 10BASE5. Typically called Thick Ethernet, or Thicknet, this
variant uses a large diameter (10mm) “thick” coaxial cable with
a 50-ohm impedance. A data rate of 10Mbps is supported
with a signaling range of 500 meters per cable segment on a
physical bus topology.
á 10BASE2. Similar to Thicknet, this variant uses a thinner
coaxial cable that can support cable runs of 185 meters. (In
this case, the “2” indicates only an approximate cable range.)
The transmission rate remains at 10Mbps, and the physical
topology is a bus. This variant typically is called Thin
Ethernet, or Thinnet.
á 10BASE-F. This variant uses fiber-optic cables to support
10Mbps signaling with a range of 4 kilometers. Three subcategories include 10BASE-FL (fiber link), 10BASE-FB
(fiber backbone), and 10BASE-FP (fiber passive).
á 10BROAD36. This broadband standard supports channel sig-
nal rates of 10Mbps. A 75-ohm coaxial cable supports cable
runs of 1,800 meters (up to 3,600 meters in a dual-cable configuration) using a physical bus topology.
NOTE
The IEEE 802.3 Physical layer definition describes signaling methods (both baseband and broadband), data rates, media, and topologies. Several Physical layer variants also have been defined. Each
variant is named following a convention that states the signaling rate
(1 or 10) in Mbps, baseband (BASE) or broadband (BROAD)
mode, and a designation of the media characteristics.
Some disagreement exists in the
industry regarding the proper use of
the name “Ethernet.” Xerox has
placed the name “Ethernet” in the
public domain, which means that no
one can claim authority over it.
Purists, however, often claim that
“Ethernet” refers to only the original
Digital-Intel-Xerox standard. More frequently, however, the term designates
any network based on CSMA/CD
access-control methods.
Usually, it is necessary to be specific
about the standard that applies to a
given network configuration. The original standard is called Ethernet version 2 (the older version 1 is still in
occasional use) or Ethernet-II. The
IEEE standard is distinguished by its
committee title as 802.3.
This distinction is important because
Ethernet version 2 and 802.3
Ethernet use incompatible frame
types. Devices using one frame type
cannot communicate with devices
using the other frame type.
87
06.919-x CH02.i
88
8/28/98 11:49 AM
C h a p te r 2
Page 88
NE TWO R KI NG S TANDARDS
á 10BASE-T. This variant uses UTP cable in a star physical
topology. The signaling rate remains at 10Mbps, and devices
can be up to 100 meters from a wiring hub.
á 100BASE-X. This proposed standard is similar to 10BASE-T
but supports 100Mbps data rates.
The IEEE 802.5 standard does not describe a cabling system. Most
implementations are based on the IBM cabling system, which uses
twisted-pair cable wired in a physical star. See Chapters 3 and 4 for
more information on Token Ring cabling and topologies.
NDIS
AND
ODI
Explain the purpose of NDIS and Novell ODI network standards.
The Network Driver Interface Specification (NDIS), a standard developed by Microsoft and 3Com Corp., describes the interface between
the network transport protocol and the Data Link layer network
adapter driver. The following list details the goals of NDIS:
á To provide a vendor-neutral boundary between the transport
protocol and the network adapter card driver so that a NDIScompliant protocol stack can operate with a NDIS-compliant
adapter driver.
á To define a method for binding multiple protocols to a single
driver so that the adapter can simultaneously support communications under multiple protocols. In addition, the method
enables you to bind one protocol to more than one adapter.
The Open Data-Link Interface (ODI), developed by Apple and
Novell, serves the same function as NDIS. Originally, ODI was written for NetWare and Macintosh environments. Like NDIS, ODI
provides rules that establish a vendor-neutral interface between the
protocol stack and the adapter driver. This interface also enables one
or more network drivers to support one or more protocol stacks.
Essentially NDIS and ODI are standards to which a person wishing
to develop a driver for a network card or a protocol will adhere.
Standards are similar to how cars are manufactured. Cars destined
for England are designed with the steering wheel on the right-hand
06.919-x CH02.i
8/28/98 11:49 AM
Page 89
C hapter 2
NETWORKING STANDARD S
side of the car. Cars for North America are designed with the steering wheel on the left-hand side of the car. This standard does not
change the function of the car or the steering wheel; conforming to
it simply ensures that the car will function properly for each country’s driving environment. NDIS and ODI are similar. Neither standard changes the function of the network card or the network card’s
driver, they simply are standards enabling the network card to function in each operating system’s environment.
C A S E S T U DY : W H AT D E V I C E S H O U L D B E U S E D : A R E P E AT E R ,
BRIDGE, OR ROUTER?
ESSENCE OF THE CASE
SCENARIO
The facts of the case are as follows:
This case study addresses the decisions involved
in determining when a LAN should implement a
repeater, bridge, or router. The company in question is a medium-size firm with roughly 100
employees. They are currently running a LAN
utilizing a single cable segment. The users of
this LAN are finding that it is taking a relatively
long time to download files from one of their two
centralized file servers. This company has two
distinct groups: the sales group and the support
group. Each group accesses its own server. Due
to the long waits, the company has called you in
to explore various options for speeding up the
throughput on the LAN. The owners of the company have heard of the terms “repeater,”
“bridge,” and “router,” and say that their LAN
supplier has mentioned that their company
should purchase all three devices.
• Too much traffic on the network is the
problem, slowing access and bringing
the productivity of the workers down.
• You must be able to define the purpose
or function of a repeater, bridge, and
router if you are to see which device is a
solution for the problem.
• To be able explain why you are proposing
which device to implement, you need to
understand at which layer in the OSI
model different components or services
run.
• You want to make sure that the solution
you are going to provide to the company
is not going to become obsolete with the
advent of new technology.
A N A LY S I S
This case study is real-world applicable, but also
puts into perspective the following two exam
objectives:
continues
89
06.919-x CH02.i
90
8/28/98 11:49 AM
C h a p te r 2
Page 90
NE TWO R KI NG S TANDARDS
C A S E S T U DY : W H AT D E V I C E S H O U L D B E U S E D : A R E P E AT E R ,
BRIDGE, OR ROUTER?
continued
• Define the communication devices that
communicate at each level of the OSI
model.
• Describe the characteristics and purpose of
the media used in IEEE 802.3 and IEEE
802.5.
The first thing to analyze is which of the three
devices can be used to isolate the network
traffic.
The repeater, which operates at the physical level
of the OSI model, is definitely ruled out. You do
so because a repeater is used to regenerate a
signal that has degenerated over distance. The
repeater has nothing to do with isolating or lessening network traffic.
Both a bridge, which operates at the Data Link
level, and a router, which operates at the
Network level of the OSI model, could be used.
As was discussed in the Data Link section, a
bridge can isolate network traffic on a single network segment. By putting a bridge on the LAN,
and placing the sales group on one side and the
support group on the other, with their respective
servers on each side, the problems with
response time brought on by network traffic
should be alleviated. That is, network traffic from
one group will not interfere with another group.
A router could also isolate network traffic, but
this would require dividing the network into two
separately addressed network segments (one
for sales and one for the support group). Recall
that routers operate at the Network layer of the
OSI model. They utilize network addresses to
determine the location of a network segment on
the network. Because routers must evaluate
packets based upon network addresses, data
transferred from one network segment to another incurs a small amount of overhead. Also,
because routers need to be programmed to be
aware of the different segment (either by using a
routing protocol or manually), the network’s
administrator requires a higher level of expertise
than that required to administer a bridge. Not all
transport protocols are routable (see Chapter 7),
and routers normally do not forward broadcasts,
thus these conditions may play a role in determining whether a router is a feasible option.
Also, routers usually cost more than bridges, and
thus are best suited for networks with more than
two segments. On this basis alone, the best
option here is that of the bridge. It provides the
solution that your client is looking for—that of
faster access—at a lower cost than a router. It
also requires less administration and is slightly
more efficient in this situation.
One issue in the “Essence of the Case” section
asked you to make sure that the technology you
are purchasing does not become obsolete. To
check for potential obsolescence, go to the IEEE
web site and look up the developments of the
various standards being investigated by the IEEE
in one of its workgroups. The type of network
this company uses (that is, Ethernet, Token Ring
or ARCNet) is not mentioned in this case study.
But this information is important to know,
because it affects which 802 standard you need
to research.
06.919-x CH02.i
8/28/98 11:49 AM
Page 91
C hapter 2
NETWORKING STANDARD S
CHAPTER SUMMARY
This chapter discussed some of the important standards that define
the networking environment. An understanding of these standards is
essential for understanding the networking topics discussed in later
chapters. This chapter covered the following:
á The OSI model
á SLIP and PPP
á The IEEE 802 Standards
á NDIS and ODI
Later chapters in this book look more closely at related topics,
including transmission media, topologies, and connectivity devices.
KEY TERMS
• International Standards
Organization (ISO)
• Open Systems Interconnection
(OSI)
• Physical layer
• Data Link layer
• Network layer
• Transport layer
• Session layer
• Presentation layer
• Application layer
• Serial Line Internet Protocol
(SLIP)
• Point-to-Point Protocol (PPP)
• IEEE
• IEEE 802 family
• circuit switching
• message switching
• packet switching
• datagram packet switching
• virtual circuit packet switching
• connection-oriented mode
• connectionless mode
• gateway
• Network Data Link Interface
Standard (NDIS)
• Open Data Link Interface (ODI)
91
06.919-x CH02.i
92
8/28/98 11:49 AM
C h a p te r 2
Page 92
NE TWO R KI NG S TANDARDS
A P P LY Y O U R L E A R N I N G
Exercises
2.1
SLIP and PPP in Dial-Up Networking
Objective: Explore the Dial-Up Networking application and learn how to configure Dial-Up Networking
for SLIP or PPP.
Estimated time: 10 minutes
1. From the Windows NT Start menu, choose
Programs/Administrative Tools and select
Dial-Up Networking. The Dial-Up Networking
application enables you to connect to another
computer as a dial-up client using SLIP or PPP.
To get the full effect of this exercise, RAS,
TCP/IP, NWLink, and NetBEUI must be
installed on your system and must be enabled for
dial-out connections. If they already are, proceed
to Step 7. Exercise 7.1 in Chapter 7 describes
how to install protocols. To enable the protocols
for dial-out connections, follow these steps:
2. Start the Control Panel Network application and
select the Services tab. If Remote Access Service
isn’t installed, click the Add button to install it.
3. Double-click on Remote Access Service in the
Services tab (or select Remote Access Service and
click Properties).
4. In the Remote Access Setup dialog box, click the
Configure button and make sure either the Dial
Out Only or Dial Out And Receive Calls button
is selected under Port Usage. Click OK.
5. In the Remote Access Setup dialog box, click the
Network button. Select dial out protocols
NetBEUI, TCP/IP, and IPX. Click OK.
6. Click Continue in the Remote Access Setup dialog box.
7. In the Dial-Up Networking main screen, click
the New button to set up a new connection. The
New Phonebook Entry dialog box appears. The
tabs of the New Phonebook Entry dialog box
enable you to enter a phone number, modem
information, security information, and a login
script. In addition, you can enter information
about the dial-up server to which you are connecting. Click the Server tab when you are finished.
8. In the Dial-Up Networking Server tab, click the
arrow to the right of the box labeled Dial-up
server type. Note that the default option is PPP:
Windows NT, Windows 95 Plus, Internet. PPP
enables you to connect to a Windows NT RAS
server, a Windows 95 machine with the Windows
95 Plus Dial-up Server feature, or a server with
an Internet-style TCP/IP configuration.
9. In the Dial-up server type box, select the SLIP:
Internet option. (TCP/IP must be installed on
your machine.) Examine the rest of the Server tab
options. The other protocols (IPX/SPX and
NetBEUI) should be grayed out, as should software compression. Your only protocol option is
TCP/IP. Check the TCP/IP check box and click
the TCP/IP Settings button. Note the boxes for a
static IP address and static DNS and WINS server addresses. Click Cancel.
10. In the New Phonebook Entry Server tab, click
the down arrow to the right of the Dial-up server
type box and choose PPP: Windows NT,
Windows 95 Plus, Internet. Note that the
IPX/SPX Compatible and NetBEUI protocol
options are now available (if they are installed on
your system and enabled for dial-out connections—see preceding note), as are software compression and PPP LCP extensions. Select the
06.919-x CH02.i
8/28/98 11:49 AM
Page 93
C hapter 2
NETWORKING STANDARD S
93
A P P LY Y O U R L E A R N I N G
TCP/IP protocol and click the TCP/IP Settings
button. Note that under a PPP connection, the
TCP/IP Settings dialog box contains option buttons for a server-assigned IP address and serverassigned name server addresses.
11. Click Cancel to exit TCP/IP Settings, Cancel to
exit New Phonebook Entry, and Close to exit
Dial-Up Networking.
2.2
Finding IEEE Standards on the Internet
Objective: Find information on the Internet about the
emergence of new standards dealing with the 802 series
of standards.
Estimated time: 30 minutes
1. On a computer that is connected to the Internet,
open up your web browser.
2. In the Location pane, type in the following URL:
www.ieee.org
3. On the welcome screen, click on the area of the
graphic that says Standards. (Because web pages
are continually being redeveloped, there may not
be a “standards” area to click on the web page. If
this is the case, do a search on the web page for
802).
Review Questions
1. Explain the difference between a router, a
repeater, and a bridge.
2. What are the seven layers of the OSI model?
3. What are the three main purposes of NDIS and
ODI?
4. What are three main differences between PPP
and SLIP?
5. In one sentence, explain the difference between
standards 802.3 and 802.5.
Exam Questions
1. The OSI model organizes communication protocols into how many layers?
A. 3
B. 7
C. 17
D. 56
2. The layers of the OSI model (in order) are
included in which of the following choices?
4. On the new window, scroll down until you see a
standards list referencing 802, LAN, and MAN
(the exact wording may change at any given time,
so looking for these key words should be directive
enough) and click on that topic.
A. Physical, Data Link, Network, Transport,
System, Presentation, Application
5. On the list of 802 series of standards, explore
what is new for the 802.3 and 802.5 set of standards.
C. Physical, Data Link, Network, Transform,
Session, Presentation, Application
B. Physical, Data Link, Network, Transport,
Session, Presentation, Application
D. Presentation, Data Link, Network, Transport,
Session, Physical, Application
06.919-x CH02.i
94
8/28/98 11:49 AM
C h a p te r 2
Page 94
NE TWO R KI NG S TANDARDS
A P P LY Y O U R L E A R N I N G
3. In the OSI model, what is the relationship of a
layer (N) to the layer above it (layer N+1)?
C. Session
D. Network
A. Layer N provides services for layer N+1.
B. Layer N+1 adds a header to information
received from layer N.
C. Layer N utilizes services provided by layer
N+1.
D. Layer N has no effect on layer N+1.
4. Which option best describes the condition of two
different computer types that can communicate?
A. They conform to the OSI model.
7. Which switching method employs virtual circuits?
A. Message
B. Circuit
C. Packet
D. All the above
8. Which OSI layer is concerned with data encryption?
B. They are both using TCP/IP.
A. Network
C. They are using compatible protocol stacks.
B. Transport
D. They are a Macintosh and a UNIX workstation.
C. Session
5. Which three of the following statements regarding protocol stacks are true?
A. A given protocol stack can run on only one
computer type.
B. Layers add headers to packets received from
higher layers in the protocol stack.
C. A protocol stack is a hierarchical set of protocols.
D. Each layer provides services for the next highest layer.
6. Which protocol layer enables multiple devices to
share the transmission medium?
A. Physical
B. Data Link
D. Presentation
9. Which switching method makes the most efficient use of network bandwidth?
A. Message
B. Circuit
C. Packet
D. All methods are about equal
10. What is another name for a message-switching
network?
A. Connectionless
B. Datagram
C. Store-and-forward
D. Virtual circuit
06.919-x CH02.i
8/28/98 11:49 AM
Page 95
C hapter 2
NETWORKING STANDARD S
95
A P P LY Y O U R L E A R N I N G
11. Which two statements about virtual circuits are
true?
A. They are usually associated with connectionoriented services.
B. A virtual circuit represents a specific path
through the network.
C. A virtual circuit appears to the connected
devices as a dedicated network path.
D. Virtual circuits dedicate a communication
channel to a single conversation.
12. Which switching method fragments messages
into small units that are routed through independent paths?
C. Session
D. Presentation
15. Which three of the following are functions of session administration?
A. Connection establishment
B. Checksum error detection
C. Data transfer
D. Connection release
16. Which two of the following are functions of connection establishment?
A. Resumption of interrupted communication
A. Message
B. Verification of logon name and password
B. Packet
C. Determination of required services
C. Circuit
D. Acknowledgment of data receipt
D. Neural
13. Which two of the following methods of dialog
control provide two-way communication?
17. Which two of the following are possible functions of the Presentation layer?
A. Data encryption
A. Simple duplex
B. Presentation of data on display devices
B. Simplex
C. Data translation
C. Half-duplex
D. Display format conversion
D. Full-duplex
14. Dialog control is a function of which layer of the
OSI reference model?
18. Which three of the following are possible functions of the Application layer?
A. Network printing service
A. Network
B. End-user applications
B. Transport
C. Client access to network services
D. Service advertisement
06.919-x CH02.i
96
8/28/98 11:49 AM
C h a p te r 2
Page 96
NE TWO R KI NG S TANDARDS
A P P LY Y O U R L E A R N I N G
19. PPP operates at which two of the following OSI
layers?
A. Physical
B. Network transport protocol
C. Physical layer
D. Network adapter driver
B. Data Link
C. Network
D. Transport
24. Routers operate at which layer of the OSI model?
A. Transport
B. Network
20. SLIP supports which of the following transport
protocols?
A. IPX/SPX
B. NetBEUI
C. TCP/IP
D. All the above
C. Data Link
D. Physical
25. Which type of communication provides flow
control at internal nodes?
A. Transport
B. Internal
21. IEEE 802.3 is associated with which of the following network architectures?
A. Token Ring
B. Ethernet
C. Internet
D. None of the above
22. IEEE 802.5 is associated with which of the following network architectures?
A. Token ring
C. Connection-oriented
D. Internet
26. Which answer best describes support over serial
line communication under the TCP/IP transport
protocol?
A. SLIP
B. PPP
C. Both A and B
D. None of the above
B. Ethernet
C. Internet
D. None of the above
23. NDIS describes the interface between which two
components?
A. User
27. 10BASE-T networks are defined in which standard?
A. IEEE 802.1
B. IEEE 802.5
C. Both A and B
D. None of the above
06.919-x CH02.i
8/28/98 11:49 AM
Page 97
C hapter 2
NETWORKING STANDARD S
97
A P P LY Y O U R L E A R N I N G
28. You notice that on your network the response
time is very slow. You wish to solve this problem.
Primary Objective: You are using protocols that
are not routable, hence there cannot be any
changes in network addresses.
Secondary Objective: The device needs to require
no user intervention.
Secondary Objective: The device needs to operate
at the Data Link layer.
Suggested Solution: You install a router in the
middle of the network cable.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
Link layer of the OSI model. See the section
titled “Components That Operate at This
Level—Bridges.”
A router is used to route data onto different network segments. It can perform the role of a
bridge as well, but if this is the case, then it is
called a brouter. A router operates at the Network
level of the OSI model. See the section titled
“Components That Operate at This Level—
Routers.”
2. The seven layers of the OSI model are as follows:
• Physical
• Data Link
• Network
• Transport
C. This solution meets the primary objective.
• Session
D. This solution does not meet the primary
objective.
• Presentation
Answers to Review Questions
1. A repeater is used to regenerate a signal. It is used
primarily to extend a cable length beyond its recommended capacity. A repeater does not route
information onto other network segments, or isolate traffic on a cable segment. A repeater operates at the Physical layer of the OSI model. See
the section titled “Components That Operate at
This Level—Repeaters.”
A bridge is responsible for isolating traffic within
a given cable segment. Some bridges can also do
the task of a repeater and regenerate a signal on a
cable. A bridge does not route data onto different
network segments. A bridge operates at the Data
• Application
See the section titled “The OSI Reference
Model.”
3. NDIS is a standard developed by Microsoft and
3Com Corp. ODI has the same function, but it
was developed by Novell and Apple. The three
main features are as follows:
• It was designed to provide a vendor-neutral
boundary between the transport protocol and
the network adapter card driver.
• It enables one network card to support multiple protocol stacks.
• It enables one protocol stack to be shared by
multiple network cards.
See the section titled “NDIS and ODI.”
06.919-x CH02.i
98
8/28/98 11:49 AM
C h a p te r 2
Page 98
NE TWO R KI NG S TANDARDS
A P P LY Y O U R L E A R N I N G
4. SLIP and PPP are designed to support dial-up
networking. PPP is more advanced than SLIP,
because the SLIP protocol has no rigid standardization. Windows NT cannot act as a SLIP server.
PPP has all the capabilities of SLIP, but also provides the following:
• Security using password logon
• Simultaneous support for multiple protocols
on the same link
• Dynamic IP addressing
• Improved error control
See the section titled “Serial Line Internet
Protocol (SLIP) and Point-to-Point Protocol
(PPP).”
5. 802.3 is an Ethernet standard, whereas 802.5 is a
Token-Ring standard.
See the section titled “The IEEE 802 Family.”
Answers to Exam Questions
1. B. The OSI model has only seven layers. See the
section titled “The OSI Reference Model.”
2. B. Answers A and C each mention a non-existent
layer (System and Transform). D references the
OSI model in the incorrect order. See the section
titled “The OSI Reference Model.”
communicate. Answers A, B, and D do not
describe situations where total conformity exists.
See the section titled “The OSI Reference
Model.”
5. B, C, D. A is incorrect because a given protocol
stack can run on many different computer types.
See the section titled “Protocol Stacks.”
6. B. Transmission media sharing is done at the
Data Link layer of the OSI model. See the section titled “OSI Data Link Layer Concepts.”
7. C. Only packet switching employs virtual circuits. See the section titled “Delivering Packets.”
8. D. Data encryption is a function of the
Presentation level. See the section titled “OSI
Presentation Layer Concepts.”
9. C. Packet switching can make the best use of network bandwidth. See the section titled “Packet
Switching.”
10. C. The other three answers do not relate to message switching. Store-and-forward is a common
term that is used to describe message-switching.
See the section titled “Message Switching.”
11. A, C. Answers B and D are the opposite of virtual circuits. See the section titled “Virtual Circuit
Packet Switching.”
12. B. A does not use small packets. C uses only one
path. D has to do with physiology. See the section titled “Delivering Packets.”
3. A. A lower layer in the OSI model provides services to the layer above it. See the section titled
“How Peer OSI Layers Communicate.”
13. C, D. A and B provide only one-way communication. See the section titled “OSI Session Layer
Concepts.”
4. C. All layers in the OSI model must be compatible with one another for two computers to
14. C. This is one of the functions of the Session
layer. See the section titled “OSI Session Layer
Concepts.”
06.919-x CH02.i
8/28/98 11:49 AM
Page 99
C hapter 2
NETWORKING STANDARD S
99
A P P LY Y O U R L E A R N I N G
15. A, C, D. Checksum error detection is done at the
Data Link layer of the OSI model. See the section titled “OSI Session Layer Concepts.”
16. B, C. Session establishment is at the start of a
session. A is in the middle of a session. D is during a session. See the section titled “OSI Session
Layer Concepts.”
17. A, C. B and D are functions of operating systems, and are not part of the OSI model. See the
section titled “OSI Presentation Layer Concepts.”
18. A, C, D. Application layer has to do with services, not desktop applications. See the section
titled “OSI Application Layer Concepts.”
19. A, B. PPP is a lower-layer protocol that encompasses both the Physical and Data Link layers of
the OSI model. See the section titled “Serial Line
Internet Protocol (SLIP) and Point-to-Point
Protocol (PPP).”
20. C. Only PPP supports all; SLIP supports only
TCP/IP. See the section titled “Serial Line
Internet Protocol (SLIP) and Point-to-Point
Protocol (PPP).”
21. B. 802.3 is an IEEE Ethernet standard. See the
section titled “IEEE 802.3.”
22. A. 802.5 is an IEEE Token-Ring standard. See
the section titled “IEEE 802.5.”
23. B, D. NDIS is a standard to which network card
drivers should be written. The NDIS standard
addresses the Data Link layer of the OSI model.
See the section titled “NDIS and ODI.”
24. B. This is the layer at which routers operate,
because routers are concerned with forwarding
packets to devices on different networks based on
each packet’s network address. See the section
titled “Components That Operate at This
Level—Routers.”
25. C. One purpose of connection-oriented communications is flow control. All the other answers are
not types of communications describing flow
control. See the section titled “OSI Session Layer
Concepts.”
26. C. Both SLIP and PPP provide TCP/IP support.
PPP also provides support for NetBEUI and
NWLink as well, whereas SLIP does not. See the
section titled “Serial Line Internet Protocol
(SLIP) and Point-to-Point Protocol (PPP).”
27. D. It is defined in the 802.3 standard. See the
section titled “IEEE 802.3.”
28. D. Adding a router accomplishes none of the
objectives. A router needs to supply more than
one network address. A router requires user input
to program network addresses into it. A router
operates at the Network Layer of the OSI model.
Suggested Readings and Resources
1. Tanenbaum, Andrew. Computer Networks.
Prentice Hall, 1996.
2. Henshall, John. Open Up OSI: An Illustrated
Guide. Ellis Horwood Ltd., 1993.
3. Simon, Alan, Tom Wheeler, and Thomas
Wheeler. Open Systems Handbook. AP
Professional, 1994.
06.919-x CH02.i
8/28/98 11:49 AM
Page 100
07.919-x Part2
8/28/98 11:56 AM
Page 101
P A R T
PLANNING
3 Transmission Media
4 Network Topologies and Architectures
5 Network Adapter Cards
6 Connectivity Devices and Transfer Mechanisms
7 Transport Protocols
8 Connection Services
9 Disaster Recovery
II
07.919-x Part2
8/28/98 11:56 AM
Page 102
08.919-x CH03.i
8/28/98 11:59 AM
Page 103
OBJECTIVES
Chapter 3 targets the first objective in the Planning
section of the Networking Essentials exam:
Select the appropriate media for various situations. Media choices include twisted-pair cable,
coaxial cable, fiber-optic cable, and wireless.
Situational elements include cost, distance limitations, and number of nodes.
. This chapter focuses on one exam objective and the
many issues that stem from it. This is due to
amount and complexity of the material associated
with the topic of Transmission Media. It is just as
important to know the advantages and disadvantages of different transmission media and in what
situations to use them as it is to simply understand
the characteristics of each transmission medium.
C H A P T E R
3
Transmission Media
08.919-x CH03.i
8/28/98 11:59 AM
Page 104
OUTLINE
Transmission Frequencies
106
Transmission Media Characteristics
108
Cost
109
Installation Requirements
109
Bandwidth
110
Band Usage (Baseband or Broadband)
Multiplexing
110
111
Attenuation
113
Electromagnetic Interference
114
Cable Media
114
Coaxial Cable
Types of Coaxial Cable
Coaxial Characteristics
Connectors for Coaxial Cable
Coax and Fire Code Classifications
114
115
116
118
120
Twisted-Pair Cable
Shielded Twisted-Pair (STP) Cable
Unshielded Twisted-Pair (UTP) Cable
121
122
125
Fiber-Optic Cable
Fiber-Optic Characteristics
128
129
Summary of Cable Characteristics
130
IBM Cabling
131
Wireless Media
132
Reasons for Wireless Networks
133
Wireless Communications with LANs
Infrared Transmission
Laser Transmission
Narrow-Band Radio Transmission
Spread-Spectrum Radio
Transmission
Microwave
133
134
135
135
135
138
Comparisons of Different Wireless
Media
141
Chapter Summary
144
08.919-x CH03.i
8/28/98 11:59 AM
Page 105
S T U DY S T R AT E G I E S
. Be able to compare and contrast one transmission media with another. Always think in terms
of which form of transmission media best suits
the needs of the network. These needs take
into account the following:
• The distances the network needs to cover
• The type of electromagnetic interference
that needs to be overcome
• The impenetrable barriers that may force you
to use a wireless media
• The relative costs of the transmission media
and whether or not they are justifiable
. Read the chapter with these criteria in mind
and pay particular attention to the tables presented. This provides you with a solid understanding of the topic of Transmission Media as
it is addressed on the exam.
08.919-x CH03.i
106
8/28/98 11:59 AM
C h a p te r 3
Page 106
TR A NS M I S S I O N MEDIA
INTRODUCTION
On any network, the various entities must communicate through
some form of media. Human communication requires some sort of
media, whether it is technologically based (as are telephone wires) or
whether it simply involves the use of our senses to detect sound
waves propagating through the air. Likewise, computers can communicate through cables, light, and radio waves. Transmission media
enable computers to send and receive messages but, as in human
communication, do not guarantee that the messages will be understood.
This chapter discusses some of the most common network transmission media. One broad classification of this transmission media is
known as bounded media, or cable media. This includes cable types
such as coaxial cable, shielded twisted-pair cable, unshielded twistedpair cable, and fiber-optic cable. Another type of media is known as
boundless media; these media include all forms of wireless communications. To lay the groundwork for these issues, the chapter begins
with an introduction to the frequencies in the electromagnetic spectrum and a look at some important characteristics of the transmission media that utilize these different frequencies to transmit the
data.
TRANSMISSION FREQUENCIES
Transmission media make possible the transmission of the electronic
signals from one computer to another. These electronic signals
express data values in the form of binary (on/off ) impulses, which
are the basis for all computer information (represented as 1s and 0s).
These signals are transmitted between the devices on the network,
using some form of transmission media (such as cables or radio)
until they reach the desired destination computer.
All signals transmitted between computers consist of some form of
electromagnetic (EM) waveform, ranging from radio frequencies
through microwaves and infrared light. Different media are used to
transmit the signals, depending on the frequency of the EM waveform. Figure 3.1 illustrates the range of electromagnetic waveforms
(known as the electromagnetic spectrum) and their associated frequencies.
08.919-x CH03.i
8/28/98 11:59 AM
Page 107
C hapter 3
FIGURE 3.1
Gamma Rays
The electromagnetic spectrum.
X-Rays
Ultraviolet
Visible Light
TeraHertz
Infrared
Extremely High
Frequency (EHF)
Super High
Frequency (SHF)
GigaHertz
Ultra High
Frequency (UHF)
Very High
Frequency (VHF)
High Frequency
(HF)
Medium Frequency
(MF)
MegaHertz
Low Frequency
(LF)
Very Low
Frequency (VLF)
Voice Frequency
(VF)
KiloHertz
Extremely Low
Frequency (ELF)
TRANSMISSION ME D IA
}
}}
}
Microwaves
RadioWaves
Audio Frequencies
≈ 30Hz - 20Khz
Power and Telephone
The electromagnetic spectrum consists of several categories of waveforms, including radio frequency waves, microwave transmissions,
and infrared light.
The frequency of a wave is dependent upon the number of waves or
oscillations that occur during a period of time. An example that all
people can relate to is the difference between a high-pitched sound,
such as a whistle, and a low-pitch sound such as a fog horn. A highpitched sound has a very high frequency; in other words, numerous
cycles of oscillation (or waves) occur each second. Whereas, a low
frequency sound, such as the fog horn, is based on relatively few
cycles or waves per second (see Figure 3.2). Although sound is not
an example of electromagnetic energy (it’s mechanical energy), the
principles are similar.
107
08.919-x CH03.i
108
8/28/98 11:59 AM
C h a p te r 3
Page 108
TR A NS M I S S I O N MEDIA
High frequency
FIGURE 3.2
High frequency and low frequency waves.
Low frequency
Radio frequency waves are often used for LAN signaling. Radio frequencies can be transmitted across electrical cables (twisted-pair or
coaxial) or by radio broadcast.
Microwave transmissions can be used for tightly focused transmissions between two points. Microwaves are used to communicate
between earth stations and satellites, for example, and they are also
used for line-of-sight transmissions on the earth’s surface. In addition, microwaves can be used in low-power forms to broadcast signals from a transmitter to many receivers. Cellular phone networks
are examples of systems that use low-power microwave signals to
broadcast signals.
Infrared light is ideal for many types of network communications.
Infrared light can be transmitted across relatively short distances and
can be either beamed between two points or broadcast from one
point to many receivers. Infrared and higher frequencies of light also
can be transmitted through fiber-optic cables. A typical television
remote control uses infrared transmission.
The next sections examine the major factors you should consider
when evaluating what type of transmission media should be implemented.
TRANSMISSION MEDIA
CHARACTERISTICS
Each type of transmission media has special characteristics that make
it suitable for a specific type of service. You should be familiar with
these characteristics for each type of media:
08.919-x CH03.i
8/28/98 11:59 AM
Page 109
C hapter 3
á Cost
á Installation requirements
á Bandwidth
á Band usage (baseband or broadband)
á Attenuation
á Immunity from electromagnetic interference
These characteristics are all important. When you design a network
for a company, all these factors play a role in the decision concerning
what type of transmission media should be used.
Cost
One main factor in the purchase decision of any networking component is the cost. Often the fastest and most robust transmission
media is desired, but a network designer must often settle for something that is slower and less robust, because it more than suffices for
the business solution at hand. The major deciding factor is almost
always price. It is a rare occasion in the field that the sky is the limit
for installing a network. As with nearly everything else in the computer field, the fastest technology is the newest, and the newest is the
most expensive. Over time, economies of scale bring the price down,
but by then, a newer technology comes along.
Installation Requirements
Installation requirements typically involve two factors. One is that
some transmission media require skilled labor to install. Bringing in
a skilled outside technician to make changes to or replace resources
on the network can bring about undue delays and costs. The second
has to do with the actual physical layout of the network. Some types
of transmission media install more easily over areas where people are
spread out, whereas other transmission media are easier to bring to
clusters of people or a roaming user.
TRANSMISSION ME D IA
109
08.919-x CH03.i
NOTE
NOTE
110
8/28/98 11:59 AM
C h a p te r 3
Page 110
TR A NS M I S S I O N MEDIA
Using the Term “Bandwidth” The
term “bandwidth” also has another
meaning. In the communications
industry, bandwidth refers to the
range of available frequencies
between the lower frequency limit and
the upper frequency limit. Frequencies
are measured in Hertz (Hz), or cycles
per second. The bandwidth of a voice
telephone line is 400–4,000Hz, which
means that the line can transmit signals with frequencies ranging from
400 to 4,000 cycles per second.
As you know, everything in computers
is represented with 1s and 0s. We
use 1s and 0s to represent the bits in
the computer. However, be sure to
remember that transmission media is
measured in megabits per second
(Mbps), not megaBYTES per second
(MBps). The difference is eight-fold,
as there are 8 bits in a byte.
Bandwidth
In computer networking, the term bandwidth refers to the measure
of the capacity of a medium to transmit data. A medium that has a
high capacity, for example, has a high bandwidth, whereas a medium
that has limited capacity has a low bandwidth.
Bandwidth can be best explained by using water hoses as an analogy.
If a half-inch garden hose can carry water flow from a trickle up to
two gallons per minute, then that hose can be said to have a bandwidth of two gallons per minute. A four-inch fire hose, however,
might have a bandwidth that exceeds 100 gallons per minute.
Data transmission rates are frequently stated in terms of the bits that
can be transmitted per second. An Ethernet LAN theoretically can
transmit 10 million bits per second and has a bandwidth of 10
megabits per second (Mbps).
The bandwidth that a cable can accommodate is determined in part
by the cable’s length. A short cable generally can accommodate
greater bandwidth than a long cable, which is one reason all cable
designs specify maximum lengths for cable runs. Beyond those limits, the highest-frequency signals can deteriorate, and errors begin to
occur in data signals. You can see this by taking a garden hose and
snapping it up and down. You can see the waves traveling down the
hose get smaller as they get farther from your hand. This loss of the
wave’s amplitude represents attenuation, or signal degradation.
Band Usage (Baseband or Broadband)
The two ways to allocate the capacity of transmission media are with
baseband and broadband transmissions. Baseband devotes the entire
capacity of the medium to one communication channel. Broadband
enables two or more communication channels to share the bandwidth of the communications medium.
Baseband is the most common mode of operation. Most LANs function in baseband mode, for example. Baseband signaling can be
accomplished with both analog and digital signals.
Although you might not realize it, you have a great deal of experience with broadband transmissions. Consider, for example, that the
TV cable coming into your house from an antenna or a cable
08.919-x CH03.i
8/28/98 12:00 PM
Page 111
C hapter 3
TRANSMISSION ME D IA
111
provider is a broadband medium. Many television signals can share
the bandwidth of the cable because each signal is modulated using a
separately assigned frequency. You can use the television tuner to
select the frequency of the channel you want to watch.
This technique of dividing bandwidth into frequency bands is called
frequency-division multiplexing (FDM) and works only with analog
signals. Another technique, called time-division multiplexing
(TDM), supports digital signals. Both of these types of multiplexing
are discussed in the next section.
Figure 3.3 contrasts the difference between baseband and broadband
modes of operation.
Multiplexing
Multiplexing is a technique that enables broadband media to support
multiple data channels. Multiplexing makes sense under a number
of circumstances:
á When media bandwidth is costly. A high-speed leased line, such
as a T1 or T3, is expensive to lease. If the leased line has sufficient bandwidth, multiplexing can enable the same line to
carry mainframe, LAN, voice, video conferencing, and various
other data types.
á When bandwidth is idle. Many organizations have installed
fiber-optic cable that is used to only partial capacity. With the
proper equipment, a single fiber can support hundreds of
megabits—or even a gigabit or more—of data per second.
á When large amounts of data must be transmitted through low-
Bandwidth
capacity channels. Multiplexing techniques can divide the original data stream into several lower-bandwidth channels, each of
which can be transmitted through a lower-capacity medium.
The signals then can be recombined at the receiving end.
{
FIGURE 3.3
Baseband
Broadband
Baseband and broadband transmission modes.
08.919-x CH03.i
112
8/28/98 12:00 PM
C h a p te r 3
Page 112
TR A NS M I S S I O N MEDIA
Multiplexing refers to combining multiple data channels for transmission on a common medium. Demultiplexing refers to recovering
the original separate channels from a multiplexed signal.
Multiplexing and demultiplexing are performed by a multiplexer
(also called a mux), which usually has both capabilities.
Frequency-Division Multiplexing
Figure 3.4 illustrates frequency-division multiplexing (FDM). This
technique works by converting all data channels to analog form.
Each analog signal can be modulated by a separate frequency (called
a “carrier frequency”) that makes it possible to recover that signal
during the demultiplexing process. At the receiving end, the demultiplexer can select the desired carrier signal and use it to extract the
data signal for that channel.
FDM can be used in broadband LANs. (A standard for Ethernet also
exists.) One advantage of FDM is that it supports bidirectional signaling on the same cable. That is, a frequency can originate from
both ends of the transmission media at once.
Time-Division Multiplexing
Time-division multiplexing (TDM) divides a channel into time slots
that are allocated to the data streams to be transmitted, as illustrated
in Figure 3.5. If the sender and receiver agree on the time-slot
assignments, the receiver can easily recover and reconstruct the original data streams.
FIGURE 3.4
Frequency-division multiplexing.
A
B
FIGURE 3.5
Time division multiplexing steams data depending on the data’s allocated time slots.
A
C B A D C B A D C B A
B
C
C
D
D
08.919-x CH03.i
8/28/98 12:00 PM
Page 113
C hapter 3
TRANSMISSION ME D IA
113
TDM transmits the multiplexed signal in baseband mode. Interestingly, this process makes it possible to multiplex a TDM signal as
one of the data channels on an FDM system.
Conventional TDM equipment utilizes fixed time divisions and allocates time to a channel, regardless of that channel’s level of activity.
If a channel isn’t busy, its time slot isn’t being fully utilized. Because
the time divisions are programmed into the configurations of the
multiplexers, this technique is often referred to as synchronous TDM.
If using the capacity of the data medium more efficiently is important, a more sophisticated technique, statistical time-division multiplexing (StatTDM), can be used. A stat-mux uses the time-slot
technique but allocates time slots based on the traffic demand on the
individual channels, as illustrated in Figure 3.6.
Notice that Channel B is allocated more time slots than Channel A,
and that Channel C is allocated the fewest time slots. Channel D is
idle, so no slots are allocated to it. To make this procedure work,
the data transmitted for each time slot includes a control field that
identifies the channel to which the data in the time slot should be
assigned.
Attenuation
Attenuation is a measure of how much a signal weakens as it travels
through a medium, as discussed in Chapter 2. This book doesn’t discuss attenuation in formal terms, but it does address the impact of
attenuation on performance.
Attenuation is a contributing factor to why cable designs must
specify limits in the lengths of cable runs. When signal strength
falls below certain limits, the electronic equipment that receives the
signal can experience difficulty isolating the original signal from the
noise present in all electronic transmissions. The effect is exactly like
trying to tune in distant radio signals. Even if you can lock on to the
A
B
C
D
A
B A B A C B B A B
B
C
FIGURE 3.6
D
Statistical time-division multiplexing allocates
timeslots based on a channel’s traffic demand.
08.919-x CH03.i
114
8/28/98 12:00 PM
C h a p te r 3
Page 114
TR A NS M I S S I O N MEDIA
signal on your radio, the sound generally still contains more noise
than the sound for a local radio station. As mentioned in the previous chapters, repeaters are used to regenerate signals; hence one solution to deal with attenuation is to add a repeater.
Electromagnetic Interference
Electromagnetic interference (EMI) consists of outside electromagnetic
noise that distorts the signal in a medium. When you listen to an
AM radio, for example, you often hear EMI in the form of noise
caused by nearby motors or lightning. Some network media are
more susceptible to EMI than others.
Crosstalk is a special kind of interference caused by adjacent wires.
Crosstalk occurs when the signal from one wire is picked up by
another wire. You may have experienced this when talking on a
telephone and hearing another conversation going on in the
background. Crosstalk is a particularly significant problem with
computer networks because large numbers of cables often are located
close together, with minimal attention to exact placement.
NOTE
CABLE MEDIA
Mixing Media Some large networks
use combinations of media. When you
mix and match different types of
media, difficulties can arise, largely
because mixed media require a
greater level of expertise and training
on the part of the network support
staff. As the number of media types
increases, your own responsibilities
increase—when a problem arises on
the LAN, the number of areas you
must investigate increases dramatically when mixed transmission media are
involved.
For the Networking Essentials exam, you need to know how to make
decisions about network transmission media based on some of the
factors described in previous sections of this chapter. The following
sections discuss three types of network cabling media, as follows:
á Coaxial cable
á Twisted-pair cable
á Fiber-optic cable
Later in this chapter, you will learn about some of the wireless communication forms.
Coaxial Cable
Coaxial cables were the first cable types used in LANs. As shown in
Figure 3.7, coaxial cable gets its name because two conductors share
08.919-x CH03.i
8/28/98 12:00 PM
Page 115
C hapter 3
TRANSMISSION ME D IA
115
a common axis; the cable is most frequently referred to as a “coax.”
A type of coaxial cable that you may be familiar with is your television cable.
The components of a coaxial cable are as follows:
á A center conductor, although usually solid copper wire, is some-
times made of stranded wire.
á An outer conductor forms a tube surrounding the center con-
ductor. This conductor can consist of braided wires, metallic
foil, or both. The outer conductor, frequently called the shield,
serves as a ground and also protects the inner conductor from
EMI.
á An insulation layer keeps the outer conductor spaced evenly
from the inner conductor.
á A plastic encasement (jacket) protects the cable from damage.
Types of Coaxial Cable
The two basic classifications for coaxial cable are as follows:
NOTE
á Thinnet
á Thicknet
The following sections discuss Thinnet and Thicknet coaxial cabling.
Thinnet
Thinnet is a light and flexible cabling medium that is inexpensive
and easy to install. Table 3.1 illustrates some Thinnet classifications.
Impedance All coaxial cables have a
characteristic measurement called
impedance, which is measured in
ohms. Impedance is a measure of the
apparent resistance to an alternating
current. You must use a cable that
has the proper impedance in any
given situation.
Insulator
Jacket
Center
Conductor
Outer
Conductor
(Shield)
FIGURE 3.7
The structure of coaxial cable consists of four
major components.
08.919-x CH03.i
116
8/28/98 12:00 PM
C h a p te r 3
Page 116
TR A NS M I S S I O N MEDIA
Note that Thinnet falls under the RG-58 family, which has a 50ohm impedance. Thinnet is approximately .25 inches (6 mm) in
thickness.
TABLE 3.1
T H I N N E T C A B L E C L A S S I F I C AT I O N S
Cable
Description
Impedance
RG-58/U
Solid copper center
50-ohm
RG-58 A/U
Wire strand center
50-ohm
RG-58 C/U
Military version of RG-58 A/U
50-ohm
RG-59
Cable TV wire
75-ohm
RG-62
ARCnet specification
93-ohm
Thinnet cable can reliably transmit a signal for 185 meters (about
610 feet).
Thicknet
Thicknet (big surprise) is thicker than Thinnet. Thicknet coaxial
cable is approximately 0.5 inches (13 mm) in diameter. Because it is
thicker and does not bend as readily as Thinnet, Thicknet cable is
harder to work with. A thicker center core, however, means that
Thicknet can carry more signals a longer distance than Thinnet.
Thicknet can transmit a signal approximately 500 meters (1,650
feet).
Thicknet cable is sometimes called Standard Ethernet (although
other cabling types described in this chapter are used for Ethernet
also). Thicknet can be used to connect two or more small Thinnet
LANs into a larger network.
Because of its greater size, Thicknet is also more expensive than
Thinnet. However, Thicknet can be installed relatively safely outside,
running from building to building.
Coaxial Characteristics
You should be familiar with the installation, cost, bandwidth, and
EMI resistance characteristics of coaxial cable. The following sections
discuss some of the characteristics of coaxial cable.
08.919-x CH03.i
8/28/98 12:00 PM
Page 117
C hapter 3
TRANSMISSION ME D IA
117
Installation
Coaxial cable is typically installed in two configurations: daisy-chain
(from device to device—Ethernet) and star (ARCnet). The daisy
chain is shown in Figure 3.8.
The Ethernet cabling shown in the figure is an example of Thinnet,
which uses RG-58 type cable. Devices connect to the cable by
means of T-connectors. Cables are used to provide connections
T-CONNECTOR
RG-58 CABLE
TO GROUND
TO OTHER
WORKSTATIONS
TERMINATOR
BNC CONNECTOR
DIRECT
ATTACHMENT
TO T-CONNECTOR
TERMINATOR
GROUNDED
TERMINATOR
REPEATER
MINIMUM DISTANCE
.5 METER
185 METERS MAXIMUM
UP TO 30 ATTACHMENTS
FIGURE 3.8
Coaxial cable wiring configuration.
08.919-x CH03.i
118
8/28/98 12:00 PM
C h a p te r 3
Page 118
TR A NS M I S S I O N MEDIA
between T-connectors. One characteristic of this type of cabling is
that the ends of the cable run must be terminated by a special connector, called a terminator. The terminator contains a resistor that is
matched to the characteristics of the cable. The resistor prevents signals that reach the end of the cable from bouncing back and causing
interference.
Coaxial cable is reasonably easy to install because the cable is robust
and difficult to damage. In addition, connectors can be installed
with inexpensive tools and a bit of practice. The device-to-device
cabling approach can be difficult to reconfigure, however, when new
devices cannot be installed near an existing cabling path.
Cost
The coaxial cable used for Thinnet falls at the low end of the cost
spectrum, whereas Thicknet is among the more costly options.
Detailed cost comparisons are made later in this chapter in
“Summary of Cable Characteristics.”
Capacity
LANs that employ coaxial cable typically have a bandwidth between
2.5Mbps (ARCNet) and 10Mbps (Ethernet). Thicker coaxial cables
offer higher bandwidth, and the potential bandwidth of coaxial is
much higher than 10Mbps. Current LAN technologies, however,
don’t take advantage of this potential. (ARCNet and Ethernet are
discussed in greater detail in Chapter 4, “Network Topologies and
Architectures.”)
EMI Characteristics
All copper media are sensitive to EMI, although the shield in coax
makes the cable fairly resistant. Coaxial cables, however, do radiate a
portion of their signal, and electronic eavesdropping equipment can
detect this radiated signal.
Connectors for Coaxial Cable
Two types of connectors are commonly used with coaxial cable. The
most common is the British Naval Connector (BNC). Figure 3.9
depicts the characteristics of BNC connectors and Thinnet cabling.
08.919-x CH03.i
8/28/98 12:00 PM
Page 119
C hapter 3
T CONNECTOR
TRANSMISSION ME D IA
119
RG-58 CABLE
TO GROUND
TO OTHER
WORKSTATIONS
TERMINATOR
BNC CONNECTOR
Key issues involving Thinnet cabling are
á A BNC T-connector connects the network board in the PC to
the network. The T-connector attaches directly to the network
board.
á BNC cable connectors attach cable segments to the T-
connectors.
á A BNC barrel connector connects to Thinnet cables.
á Both ends of the cable must be terminated. A BNC terminator
is a special connector that includes a resistor that is carefully
matched to the characteristics of the cable system.
á One of the terminators must be grounded. A wire from the
connector is attached to a grounded point, such as the center
screw of a grounded electrical outlet.
In contrast, Thicknet uses N-connectors, which screw on rather than
use a twist lock (see Figure 3.10). As with Thinnet, both ends of the
cable must be terminated, and one end must be grounded.
Workstations don’t connect directly to the cable with Thicknet.
Instead, a connecting device called a transceiver is attached to the
Thicknet cable. This transceiver has a port for an AUI connector
(which looks deceivingly like a joystick connector), and an AUI
cable (also called a transceiver cable or a drop cable) connects the
workstation to the Thicknet medium. Transceivers can connect to
Thicknet cables in the following two ways:
FIGURE 3.9
Thinnet is connected using BNC T-connectors.
08.919-x CH03.i
120
8/28/98 12:00 PM
C h a p te r 3
N-SERIES
TERMINATOR
Page 120
TR A NS M I S S I O N MEDIA
THICK NET
TRANSCEIVER (MAU)
CLAMP-ON
TRANSCEIVER (CMAU)
N-SERIES
CONNECTOR
TO OTHER
WORKSTATIONS
TO
GROUND
AUI CONNECTOR
AUI CABLE
FIGURE 3.10
Connectors and cabling for Thicknet.
á Transceivers can be connected by cutting the cable and splicing
N-connectors and a T-connector on the transceiver. Because it
is so labor-intensive, this original method of connecting is used
rather infrequently.
NOTE
á The more common approach is to use a clamp-on transceiver,
AUI port connectors sometimes are
called DIX connectors or DB-15 connectors.
which has pins that penetrate the cable without the need for
cutting it. Because clamp-on transceivers force sharp teeth into
the cable, they frequently are referred to as vampire taps.
You can use a transceiver to connect a Thinnet LAN to a Thicknet
backbone.
Coax and Fire Code Classifications
The space above a drop ceiling (between the ceiling and the floor of a
building’s next level) is extremely significant to both network administrators and fire marshals. This space (called the plenum—see Figure
3.11) is a convenient place to run network cables around a building.
The plenum, however, is typically an open space in which air circulates freely, and, consequently, fire marshals pay special attention to it.
The most common outer covering for coaxial cabling is polyvinyl
chloride (PVC). PVC cabling gives off poisonous fumes when it
burns. For that reason, fire codes prohibit PVC cabling in the
08.919-x CH03.i
8/28/98 12:00 PM
Page 121
C hapter 3
TRANSMISSION ME D IA
121
Plenum
plenum because poisonous fumes in the plenum can circulate freely
throughout the building.
Plenum-grade coaxial cabling is specially designed to be used without
conduit in plenums, walls, and other areas where fire codes prohibit
PVC cabling. Plenum-grade cabling is less flexible and more expensive than PVC cabling, so it is used primarily where PVC cabling
can’t be used.
Twisted-Pair Cable
Twisted-pair cable has become the dominant cable type for all new
network designs that employ copper cable. Among the several reasons for the popularity of twisted-pair cable, the most significant is
its low cost. Twisted-pair cable is inexpensive to install and offers the
lowest cost per foot of any cable type. Your telephone cable is an
example of a twisted-pair type cable.
A basic twisted-pair cable consists of two strands of copper wire
twisted together (see Figure 3.12). The twisting reduces the sensitivity of the cable to EMI and also reduces the tendency of the cable to
radiate radio frequency noise that interferes with nearby cables and
electronic components, because the radiated signals from the twisted
wires tend to cancel each other out. (Antennas, which are purposely
designed to radiate radio frequency signals, consist of parallel, not
twisted, wires).
FIGURE 3.11
The plenum—the space between the drop-down
ceiling of a room and its actual ceiling—is often
a convenient spot for placing network cabling.
08.919-x CH03.i
122
8/28/98 12:00 PM
C h a p te r 3
Page 122
TR A NS M I S S I O N MEDIA
FIGURE 3.12
Twisted-pair cabling.
Insulating
Jackets
Conductors
Twisting of the wires also controls the tendency of the wires in the
pair to cause EMI in each other. As noted previously, whenever two
wires are in close proximity, the signals in each wire tend to produce
crosstalk in the other. Twisting the wires in the pair reduces crosstalk
in much the same way that twisting reduces the tendency of the
wires to radiate EMI.
A twisted-pair cable is used in most cases to connect a PC to either a
HUB or a MAU. Both of these devices are discussed in Chapter 6,
“Connectivity Devices and Transfer Mechanisms.” Two types of
twisted-pair cable are used in LANs: shielded and unshielded, as
explained in the following section.
Shielded Twisted-Pair (STP) Cable
Shielded twisted-pair cabling consists of one or more twisted pairs of
cables enclosed in a foil wrap and woven copper shielding. Figure
3.13 shows IBM Type 1 cabling, the first cable type used with IBM
Token Ring. Early LAN designers used shielded twisted-pair cable
because the shield performed double duty, reducing the tendency of
the cable to radiate EMI and reducing the cable’s sensitivity to outside interference.
Jacket
Shield
FIGURE 3.13
A shielded twisted-pair cable.
Two
Twisted
Pairs
Coaxial and STP cables use shields for the same purpose. The shield
is connected to the ground portion of the electronic device to which
the cable is connected. A ground is a portion of the device that
serves as an electrical reference point, and usually, it is literally connected to a metal stake driven into the ground. A properly grounded
shield prevents signals from getting into or out of the cable.
The picture in Figure 3.13 is an example of IBM Type 1 cable, an
STP cable, and includes two twisted pairs of wire within a single
shield. Various types of STP cable exist, some that shield each pair
individually and others that shield several pairs. The engineers who
design a network’s cabling system choose the exact configuration.
IBM designates several twisted-pair cable types to use with their
Token Ring network design, and each cable type is appropriate for a
08.919-x CH03.i
8/28/98 12:00 PM
Page 123
C hapter 3
TRANSMISSION ME D IA
123
given kind of installation. A completely different type of STP is the
standard cable for Apple’s AppleTalk network.
Because so many different types of STP cable exist, describing precise characteristics is difficult. The following sections, however, offer
some general guidelines.
Cost
STP cable costs more than thin coaxial or unshielded twisted-pair
cable. STP is less costly, however, than thick coax or fiber-optic
cable.
Installation
Naturally, different network types have different installation requirements. One major difference is the connector used. Apple LocalTalk
connectors generally must be soldered during installation, a process
that requires some practice and skill on the part of the installer. IBM
Token Ring uses a so called unisex or hermaphrodite data connector
(the connectors are both male and female), which can be installed
with such common tools as a knife, a wire stripper, and a large pair
of pliers (see Figure 3.14).
In many cases, installation can be greatly simplified with prewired
cables—cables precut to length and installed with the appropriate
connectors. You must learn to install the required connectors, however, when your installation requires the use of bulk cable. The
installation of cables has been regulated or made part of building
codes in some areas, to be performed only by a certified cable
FIGURE 3.14
An IBM Data connector, also known as a hermaphrodite connector.
08.919-x CH03.i
124
8/28/98 12:01 PM
C h a p te r 3
Page 124
TR A NS M I S S I O N MEDIA
installer. You should check the regulations regarding this in your area
before beginning the installation of any cable.
CONNE CTOR TY P E S
Most connectors require two connector types to complete a connection. The traditional designation for connector types is male and
female. The male connector is the connector with pins, and the
female connector has receptacles into which the pins insert. In a
standard AC wall outlet, for example, the outlet itself is female and
the plug on the line cord is male.
These designations originated when electrical installation was a
male province so the terms “male” and “female” are being
replaced gradually. A commonly used alternative is “pins and
sockets.”
The IBM data connector is called a unisex or hermaphrodite connector because the connector has both pins and sockets. Any IBM
data connector can connect to any other IBM data connector.
STP cable tends to be rather bulky. IBM Type 1 cable is approximately 1⁄2 inch (13 mm) in diameter. Therefore, cable paths cannot
hold nearly as many STP cables as they can when a thinner medium
is used.
Capacity
STP cable has a theoretical capacity of 500Mbps, although few
implementations exceed 155Mbps with 100-meter cable runs. The
most common data rate for STP cable is 16Mbps, which is the top
data rate for Token Ring networks.
Attenuation
All varieties of twisted-pair cable have attenuation characteristics that
limit the length of cable runs to a few hundred meters, although a
100-meter limit is most common.
EMI Characteristics
The shield in STP cable results in good EMI characteristics for copper cable, comparable to the EMI characteristics of coaxial cable.
This is one reason STP might be preferred to unshielded twisted-
08.919-x CH03.i
8/28/98 12:01 PM
Page 125
C hapter 3
TRANSMISSION ME D IA
125
pair cable in some situations. As with all copper cables, STP is still
sensitive to interference and vulnerable to electronic eavesdropping.
Connectors for STP
AppleTalk and Token Ring networks can be cabled using UTP cable
and RJ-45 connectors (described later in this chapter), but both networks originated as STP cabling systems. For STP cable, AppleTalk
also employs a DIN-type connector. Figure 3.15 shows an IBM connector connected to a network card having a DIN (DB-9) connector
using a STP cable.
The IBM Data Connector is unusual because it doesn’t come in two
gender configurations. Instead, any IBM Data Connector can be
snapped to any other IBM Data Connector. The IBM cabling system is discussed later in this chapter.
Unshielded Twisted-Pair (UTP) Cable
Unshielded twisted-pair cable doesn’t incorporate a braided shield
into its structure. However, the characteristics of UTP are similar in
many ways to STP, differing primarily in attenuation and EMI. As
shown in Figure 3.16, several twisted pairs can be bundled together
in a single cable. These pairs are typically color-coded to distinguish
them.
DB-9 Connector
FIGURE 3.15
A drop cable using a DB-9 connector to connect
to the Network Interface Card (NIC), and having
IBM Data Connector ready to be attached to a
MAU.
Shielded Twisted-Pair Cable
IBM Data Connector
FIGURE 3.16
A multipair UTP cable.
08.919-x CH03.i
126
8/28/98 12:01 PM
C h a p te r 3
Page 126
TR A NS M I S S I O N MEDIA
Telephone systems commonly use UTP cabling. Network engineers
can sometimes use existing UTP telephone cabling (if it is new
enough and of a high enough quality to support network communications) for network cabling.
UTP cable is a latecomer to high-performance LANs because engineers only recently solved the problems of managing radiated noise
and susceptibility to EMI. Now, however, a clear trend toward UTP
is in operation, and all new copper-based cabling schemes are based
on UTP.
UTP cable is available in the following five grades, or categories:
á Categories 1 and 2. These voice-grade cables are suitable only
for voice and for low data rates (below 4Mbps). Category 1
was once the standard voice-grade cable for telephone systems.
The growing need for data-ready cabling systems, however, has
caused Categories 1 and 2 cable to be supplanted by Category
3 for new installations.
á Category 3. As the lowest data-grade cable, this type of cable
generally is suited for data rates up to 10Mbps. Some innovative schemes utilizing new standards and technologies, however, enable the cable to support data rates up to 100Mbps.
Category 3, which uses four twisted pairs with three twists per
foot, is now the standard cable used for most telephone installations.
á Category 4. This data-grade cable, which consists of four
twisted-pairs, is suitable for data rates up to 16Mbps.
á Category 5. This data-grade cable, which also consists of four
twisted-pairs, is suitable for data rates up to 100Mbps. Most
new cabling systems for 100Mbps data rates are designed
around Category 5 cable.
The price of the grades of cable increase as you move from Category
1 to Category 5.
In a UTP cabling system, the cable is only one component of the
system. All connecting devices are also graded, and the overall
cabling system supports only the data rates permitted by the lowestgrade component in the system. In other words, if you require a
Category 5 cabling system, all connectors and connecting devices
must be designed for Category 5 operation.
08.919-x CH03.i
8/28/98 12:01 PM
Page 127
C hapter 3
The installation procedures for Category 5 cable also have more
stringent requirements than the lower cable categories. Installers of
Category 5 cable require special training and skills to understand
these more rigorous requirements.
UTP cable offers an excellent balance of cost and performance characteristics, as discussed in the following sections.
Cost
UTP cable is the least costly of any cable type, although properly
installed Category 5 tends to be fairly expensive. In some cases, existing cable in buildings can be used for LANs, although you should
verify the category of the cable and know the length of the cable in
the walls. Distance limits for voice cabling are much less stringent
than for data-grade cabling.
Installation
UTP cable is easy to install. Some specialized equipment might be
required, but the equipment is low in cost and its use can be mastered with a bit of practice. Properly designed UTP cabling systems
easily can be reconfigured to meet changing requirements.
As noted earlier, however, Category 5 cable has stricter installation
requirements than lower categories of UTP. Special training is recommended for dealing with Category 5 UTP.
Capacity
The data rates possible with UTP have pushed up from 1Mbps, past
4 and 16Mbps, to the point where 100Mbps data rates are now
common.
Attenuation
UTP cable shares similar attenuation characteristics with other copper cables. UTP cable runs are limited to a few hundred meters,
with 100 meters (a little more than 300 feet) as the most frequent
limit.
EMI Characteristics
Because UTP cable lacks a shield, it is more sensitive to EMI than
coaxial or STP cables. The latest technologies make it possible to use
UTP in the vast majority of situations, provided that reasonable care
TRANSMISSION ME D IA
127
08.919-x CH03.i
128
8/28/98 12:02 PM
C h a p te r 3
Page 128
TR A NS M I S S I O N MEDIA
is taken to avoid electrically noisy devices such as motors and fluorescent lights. Nevertheless, UTP might not be suitable for noisy
environments such as factories. Crosstalk between nearby unshielded
pairs limits the maximum length of cable runs.
Connectors for UTP
The most common connector used with UTP cables is the RJ-45
connector shown in Figure 3.17. These connectors are easy to install
on cables and are also extremely easy to connect and disconnect. An
RJ-45 connector has eight pins and looks like a common RJ-11 telephone connector. They are slightly different sizes, however, and
won’t fit together: an RJ-11 has only four pins.
Distribution racks, trays, shelves, and patch panels are available for
large UTP installations. These accessories enable you to organize network cabling and also provide a central spot for expansion and
reconfiguration. One necessary accessory, a jack coupler, is a small
device that attaches to a wall plate or a patch panel and receives an
RJ-45 connection. Jack couplers can support transmission speeds of
up to 100Mbps.
Fiber-Optic Cable
In almost every way, fiber-optic cable is the ideal cable for data transmission. Not only does this type of cable accommodate extremely
high bandwidths, but it also presents no problems with EMI and
supports durable cables and cable runs as long as several kilometers.
The two disadvantages of fiber-optic cable, however, are cost and
installation difficulty. Despite these disadvantages, fiber-optic cable is
now often installed into buildings by telephone companies as the
cable of choice.
FIGURE 3.17
An RJ-45 connector.
Jacket
(Sheath)
Cladding
FIGURE 3.18
A fiber-optic cable.
Fiber
Core
The center conductor of a fiber-optic cable is a fiber that consists of
highly refined glass or plastic designed to transmit light signals with
little loss. A glass core supports a longer cabling distance, but a plastic core is typically easier to work with. The fiber is coated with a
cladding or a gel that reflects signals back into the fiber to reduce
signal loss. A plastic sheath protects the fiber (see Figure 3.18).
A fiber-optic network cable consists of two strands separately
enclosed in plastic sheaths. One strand sends and the other receives.
Two types of cable configurations are available: loose and tight
08.919-x CH03.i
8/28/98 12:02 PM
Page 129
C hapter 3
TRANSMISSION ME D IA
129
Optical fiber cables don’t transmit electrical signals. Instead, the data
signals must be converted into light signals. Light sources include
lasers and light-emitting diodes (LEDs). LEDs are inexpensive but
produce a fairly poor quality of light suitable for only less-stringent
applications.
The end of the cable that receives the light signal must convert the
signal back to an electrical form. Several types of solid-state components can perform this service.
One of the significant difficulties of installing fiber-optic cable arises
when two cables must be joined. The small cores of the two cables
(some are as small as 8.3 microns) must be lined up with extreme
precision to prevent excessive signal loss.
Fiber-Optic Characteristics
As with all cable types, fiber-optic cables have their share of advantages and disadvantages.
Cost
The cost of the cable and connectors has fallen significantly in recent
years. However, the electronic devices required are significantly more
expensive than comparable devices for copper cable. Fiber-optic
cable is also the most expensive cable type to install.
Installation
Greater skill is required to install fiber-optic cable than to install
most copper cables. Improved tools and techniques, however, have
reduced the training required. Still, fiber-optic cable requires greater
care because the cables must be treated fairly gently during installation. Every cable has a minimum bend radius, for example, and
fibers are damaged if the cables are bent too sharply. It also is important to not stretch the cable during installation.
NOTE
configurations. Loose configurations incorporate a space between the
fiber sheath and the outer plastic encasement; this space is filled with
a gel or other material. Tight configurations contain strength wires
between the conductor and the outer plastic encasement. In both
cases, the plastic encasement must supply the strength of the cable,
while the gel layer or strength wires protect the delicate fiber from
mechanical damage.
Lasers A laser is a light source that
produces an especially pure light that
is monochromatic (one color) and
coherent (all waves are parallel). The
most commonly used source of laser
light in LAN devices is called an injection laser diode (ILD). The purity of
laser light makes lasers ideally suited
to data transmissions because they
can work with long distances and high
bandwidths. Lasers, however, are
expensive light sources used only
when their special characteristics are
required.
08.919-x CH03.i
130
8/28/98 12:02 PM
C h a p te r 3
Page 130
TR A NS M I S S I O N MEDIA
Capacity
Fiber-optic cable can support high data rates (as high as
200,000Mbps) even with long cable runs. Although UTP cable runs
are limited to less than 100 meters with 100Mbps data rates, fiberoptic cables can transmit 100Mbps signals for several kilometers.
Attenuation
Attenuation in fiber-optic cables is much lower than in copper
cables. Fiber-optic cables are capable of carrying signals for several
kilometers.
EMI Characteristics
Because fiber-optic cables don’t use electrical signals to transmit data,
they are totally immune to electromagnetic interference. The cables
are also immune to a variety of electrical effects that must be taken
into account when designing copper cabling systems.
When electrical cables are connected between two buildings, the
ground potentials (voltages) between the two buildings can differ.
When a difference exists (as it frequently does), the current flows
through the grounding conductor of the cable, even though the
ground is supposed to be electrically neutral and no current should
flow. When current flows through the ground conductor of a cable,
the condition is called a ground loop. Ground loops can result in
electrical instability and various other types of anomalies. Because
fiber-optic cable is immune to electrical effects, the best way to connect networks in different buildings is by putting in a fiber-optic
link segment. Fiber-optic cable also makes a great backbone for larger networks.
Because the signals in fiber-optic cable are not electrical in nature,
they cannot be detected by the electronic eavesdropping equipment
that detects electromagnetic radiation. Therefore, fiber-optic cable is
the perfect choice for high-security networks.
Summary of Cable Characteristics
The table below summarizes the characteristics of the four cable
types discussed in this section.
08.919-x CH03.i
8/28/98 12:02 PM
Page 131
C hapter 3
C O M PA R I S O N
OF
TRANSMISSION ME D IA
R E V I E W
CABLE MEDIA
B R E A K
Cable Type
Cost
Installation
Capacity
Range
EMI
Coaxial Thinnet
Less than STP
Inexpensive/easy
10Mbps typical
185 m
Less sensitive than UTP
Coaxial Thicknet
Greater than STP,
less than fiber
Easy
10Mbps typical
500 m
Less sensitive than UTP
Shielded twistedpair (STP)
Greater than UTP,
less than Thicknet
Fairly easy
16Mbps typical
up to 500Mbps
100 m typical
Less sensitive than UTP
Unshielded
twisted-pair (UTP)
Lowest
Inexpensive/easy
10Mbps typical
up to 100Mbps
100 m typical
Most sensitive
Fiber-optic
Highest
Expensive/difficult
100Mbps typical
10s of kilometers
Insensitive
When comparing cabling types, remember that the characteristics
you observe are highly dependent on the implementations, such as
the network cards, hubs, and other devices used. Engineers once
thought that UTP cable would never reliably support data rates
above 4Mbps, but 100Mbps data rates are now common.
Some comparisons between cable types are fairly involved. For example, although fiber-optic cable is costly on a per-foot basis, it may be
the most cost-effective alternative when you need to run a cable for
many kilometers. To build a copper cable many kilometers in length,
you need to install repeaters at several points along the cable to
amplify the signal. These repeaters could easily exceed the cost of a
fiber-optic cable run.
IBM Cabling
IBM assigns separate names, standards, and specifications for network cabling and cabling components. These IBM cabling types
roughly parallel standard forms used elsewhere in the industry, as
Table 3.2 illustrates. The AWG designation in this table stands for
the American Wire Gauge standard, a specification for wire gauges.
Higher gauge wire is thinner; lower gauge wire is thicker.
IBM provides a unique connector (mentioned earlier in this chapter)
that is of both genders—any two of the same type can be connected
together. IBM also uses other types of connectors, such as the standard RJ-45 used in many office environments.
131
08.919-x CH03.i
132
8/28/98 12:02 PM
C h a p te r 3
Page 132
TR A NS M I S S I O N MEDIA
TABLE 3.2
I BM C A B L I N G T Y P E S
Cable Type
Description
Comment
Type 1
Shielded twisted-pair (STP)
Two twisted pairs of 22AWG wire in braided shield
Type 2
Voice and data
Two twisted pairs of 22AWG wire for data and braided shield, and two twisted pairs of
26AWG for voice
Type 3
Voice
Four solid UTP pairs; 22 or 24AWG wire
Type 4
Not defined
Type 5
Fiber-optic
Two 62.5/125-micron multi-mode fibers
Type 6
Data patch cable
Two twisted pairs of 26AWG wire, dual foil, and braided shield
Type 7
Not defined
Type 8
Carpet grade
Two twisted pairs of 26 AWG wire with shield for use under carpets
Type 9
Plenum grade
Two twisted pairs, shielded (see previous discussion of plenum-grade cabling)
This list of IBM cable types is important, as many shops and documentation often reference cable types using the IBM classification.
NOTE
WIRELESS MEDIA
Point-to-point Connectivity Wireless
point-to-point communications are
another facet of wireless LAN technology. Point-to-point wireless technology
specifically facilitates communications
between a pair of devices (rather than
attempting to achieve an integrated
networking capability). For instance, a
point-to-point connection might transfer data between a laptop and a
home-based computer or between a
computer and a printer. Point-to-point
signals, if powerful enough, can pass
through walls, ceilings, and other
obstructions. Point-to-point provides
data transfer rates of 1.2 to
38.4Kbps for a range of up to 200
feet indoors (or one third of a mile for
line-of-sight broadcasts).
The extraordinary convenience of wireless communications has
placed an increased emphasis on wireless networks in recent years.
Technology is expanding rapidly and will continue to expand into
the near future, offering more and better options for wireless networks.
Presently, you can subdivide wireless networking technology into
three basic types corresponding to three basic networking scenarios:
á Local area networks (LANs). Occasionally you will see a fully
wireless LAN, but more typically one or more wireless
machines function as members of a cable-based LAN.
á Extended local networks. A wireless connection serves as a back-
bone between two LANs. For instance, a company with office
networks in two nearby but separate buildings could connect
those networks using a wireless bridge.
á Mobile computing. A mobile machine connects to the home
network using cellular or satellite technology.
The following sections describe these technologies and some of the
networking options available with each.
08.919-x CH03.i
8/28/98 12:02 PM
Page 133
C hapter 3
Reasons for Wireless Networks
Wireless networks are especially useful for the following situations:
á Spaces where cabling would be impossible or inconvenient.
These include open lobbies, inaccessible parts of buildings,
older buildings, historical buildings where renovation is prohibited, and outdoor installations.
á People who move around a lot within their work environment.
Network administrators, for instance, must troubleshoot a
large office network. Nurses and doctors need to make rounds
at a hospital.
á Temporary installations. These situations include any tempo-
rary department set up for a specific purpose that soon will be
torn down or relocated.
á People who travel outside of the work environment and need
instantaneous access to network resources.
á Satellite offices or branches, ships in the ocean, or teams in
remote field locations that need to be connected to a main
office or location.
Wireless Communications with LANs
For some of the reasons described earlier in this chapter, it is often
advantageous for a network to include some wireless nodes.
Typically, though, the wireless nodes are part of what is otherwise a
traditional, cable-based network.
An access point is a stationary transceiver connected to the cablebased LAN that enables the cordless PC to communicate with the
network. The access point acts as a conduit for the wireless PC. The
process is initiated when the wireless PC sends a signal to the access
point; from there, the signal reaches the network. The truly wireless
communication, therefore, is the communication from the wireless
PC to the access point. Use of an access point transceiver is one of
several ways to achieve wireless networking. Some of the others are
described in later sections.
This is similar to when you use your remote control for your TV.
Think of the remote control unit in your hand as the computer, and
TRANSMISSION ME D IA
133
08.919-x CH03.i
134
8/28/98 12:02 PM
C h a p te r 3
Page 134
TR A NS M I S S I O N MEDIA
the area on the TV set that receives the signal as your access point,
or stationary receiver.
You can classify wireless LAN communications according to transmission method. The four most common LAN wireless transmission
methods are as follows:
á Infrared
á Laser
á Narrow-band radio
á Spread-spectrum radio
á Microwave
The following sections look briefly at these important wireless transmission methods. Because of vast differences in evaluation criteria
such as costs, ease of installation, distance, and EMI characteristics,
these items are evaluated at the end of this section in a summary
table. (Bandwidth usage is not evaluated because wireless media is
not a bound communication media.)
Infrared Transmission
You use an infrared communication system every time you control
your television with a remote control. The remote control transmits
pulses of infrared light that carry coded instructions to a receiver on
the TV. This technology also is used for network communication.
Four varieties of infrared communications are as follows:
á Broadband optical telepoint. This method uses broadband tech-
nology. Data transfer rates in this high-end option are competitive with those for a cable-based network.
á Line-of-sight infrared. Transmissions must occur over a clear
line-of-sight path between transmitter and receiver.
á Reflective infrared. Wireless PCs transmit toward a common,
central unit, which then directs communication to each of the
nodes.
á Scatter infrared. Transmissions reflect off floors, walls, and ceil-
ings until (theoretically) they finally reach the receiver. Because
of the imprecise trajectory, data transfer rates are slow. The
maximum reliable distance is around 100 feet.
08.919-x CH03.i
8/28/98 12:02 PM
Page 135
C hapter 3
TRANSMISSION ME D IA
135
Infrared transmissions are typically limited to within 100 feet.
Within this range, however, infrared is relatively fast. Infrared’s high
bandwidth supports transmission speeds of up to 10Mbps.
Infrared devices are insensitive to radio-frequency interference, but
reception can be degraded by bright light. Because transmissions are
tightly focused, they are fairly immune to electronic eavesdropping.
Infrared transmissions are commonly used for LAN transmissions,
yet can also be employed for WAN transmissions as well.
High-powered laser transmitters can transmit data for several thousand yards when line-of-sight communication is possible. Lasers can
be used in many of the same situations as microwave links (described
later in this chapter), but do not require an FCC license. On a LAN
scale, laser light technology is similar to infrared technology. Laser
light technology is employed in both LAN and WAN transmissions,
though it is more commonly used in WAN transmissions.
Narrow-Band Radio Transmission
In narrow-band radio communications (also called single-frequency
radio), transmissions occur at a single radio frequency. The range of
narrow-band radio is greater than that of infrared, effectively
enabling mobile computing over a limited area. Neither the receiver
nor the transmitter must be placed along a direct line of sight; the
signal can bounce off walls, buildings, and even the atmosphere, but
heavy walls, such as steel or concrete enclosures, can block the signal.
Spread-Spectrum Radio Transmission
Spread-spectrum radio transmission is a technique originally developed by the military to solve several communication problems.
Spread-spectrum improves reliability, reduces sensitivity to interference and jamming, and is less vulnerable to eavesdropping than
single-frequency radio. Spread-spectrum radio transmissions are
commonly used for WAN transmissions that connect multiple LANs
or network segments together.
As its name suggests, spread-spectrum transmission uses multiple frequencies to transmit messages. Two techniques employed are frequency hopping and direct sequence modulation.
NOTE
Laser Transmission
FCC License An FCC license is
required to use certain radio frequencies. Some of these reserved frequencies are the ones airline pilots and
police communications utilize.
08.919-x CH03.i
136
8/28/98 12:02 PM
C h a p te r 3
Page 136
TR A NS M I S S I O N MEDIA
Frequency hopping switches (hops) among several available frequencies (see Figure 3.19), staying on each frequency for a specified interval of time. The transmitter and receiver must remain synchronized
during a process called a “hopping sequence” for this technique to
work. Range for this type of transmission is up to two miles outdoors and 400 feet indoors. Frequency hopping typically transmits at
up to 250Kbps, although some versions can reach as high as 2Mbps.
Direct sequence modulation breaks original messages into parts
called chips (see Figure 3.20), which are transmitted on separate
frequencies. To confuse eavesdroppers, decoy data also can be transmitted on other frequencies. The intended receiver knows which
frequencies are valid and can isolate the chips and reassemble the
message. Eavesdropping is difficult because the correct frequencies
are not known, and the eavesdropper cannot isolate the frequencies
carrying true data. Because different sets of frequencies can be
selected, this technique can operate in environments that support
other transmission activity. Direct sequence modulation systems
operating at 900MHz support bandwidths of 2–6Mbps.
Spread-spectrum radio transmissions are often used to connect multiple LAN segments together, thus it is often a WAN connection.
Frequency A
10
00
11
1
11
0 0
Frequency B
0
Frequency C
FIGURE 3.19
Frequency hopping transmits data over various
frequencies for specific periods of time.
Data to transmit
111 000 110 010
08.919-x CH03.i
8/28/98 12:02 PM
Page 137
C hapter 3
TRANSMISSION ME D IA
FIGURE 3.20
0
Frequency A
1
Frequency B
0
Frequency C
0
0
1
1
0
1
1
0
1
Data to transmit
111 000 110 010
W I R E L E S S B RI DGES
Wireless technology can connect LANs in two different buildings
into an extended LAN. This capability is, of course, also available
through other technologies (such as a T1 line—discussed in
Chapter 6—or a leased line from a telephone provider), but depending on the conditions, a wireless solution is sometimes more costeffective. A wireless connection between two buildings also provides a solution to the ground potential problem described in a
note earlier in this chapter.
A wireless bridge acts as a network bridge, merging two local LANs
over a wireless connection. (See Chapter 2, “Networking
Standards,” and Chapter 6 for more information on bridges.)
Wireless bridges typically use spread-spectrum radio technology to
transmit data for up to three miles. (Antennae at each end of the
bridge should be placed in an appropriate location, such as a
rooftop.) A device called a long-range wireless bridge has a range of
up to 25 miles.
Direct sequence modulation.
137
08.919-x CH03.i
138
8/28/98 12:02 PM
C h a p te r 3
Page 138
TR A NS M I S S I O N MEDIA
Microwave
Microwave technology has applications in all three of the wireless
networking scenarios: LAN, extended LAN, and mobile networking.
As shown in Figure 3.21, microwave communication can take two
forms: terrestrial (ground) links and satellite links. The frequencies
and technologies employed by these two forms are similar, but distinct differences exist between them.
Mobile computing is a growing technology that provides almost
unlimited range for traveling computers by using satellite and cellular phone networks to relay the signal to a home network. Mobile
computing typically is used with portable PCs or personal digital
assistant (PDA) devices.
Three forms of mobile computing are as follows:
á Packet-radio networking. The mobile device sends and receives
network-style packets via satellite. Packets contain a source and
destination address, and only the destination device can receive
and read the packet.
á Cellular networking. The mobile device sends and receives
cellular digital packet data (CDPD) using cellular phone technology and the cellular phone network. Cellular networking
provides very fast communications.
Satellite
Link
Ground
Link
FIGURE 3.21
Terrestrial and satellite microwave links.
08.919-x CH03.i
8/28/98 12:02 PM
Page 139
C hapter 3
á Satellite station networking. Satellite mobile networking stations
use satellite microwave technology, which is described later in
this chapter.
Terrestrial Microwave
Terrestrial microwave communication employs earth-based transmitters and receivers. The frequencies used are in the low gigahertz
range, which limits all communications to line-of-sight. You
probably have seen terrestrial microwave equipment in the form of
telephone relay towers, which are placed every few miles to relay
telephone signals across a country.
Microwave transmissions typically use a parabolic antenna that produces a narrow, highly directional signal. A similar antenna at the
receiving site is sensitive to signals only within a narrow focus.
Because the transmitter and receiver are highly focused, they must be
adjusted carefully so that the transmitted signal is aligned with the
receiver.
A microwave link is used frequently to transmit signals in instances
in which it would be impractical to run cables. If you need to connect two networks separated by a public road, for example, you
might find that regulations restrict you from running cables above or
below the road. In such a case, a microwave link is an ideal solution.
Some LANs operate at microwave frequencies at low power and use
nondirectional transmitters and receivers. Network hubs can be
placed strategically throughout an organization, and workstations
can be mobile or fixed. This approach is one way to enable mobile
workstations in an office setting.
In many cases, terrestrial microwave uses licensed frequencies. A
license must be obtained from the FCC, and equipment must be
installed and maintained by licensed technicians.
Terrestrial microwave systems operate in the low gigahertz range,
typically at 4–6GHz and 21–23GHz, and costs are highly variable
depending on requirements. Long-distance microwave systems can
be quite expensive but might be less costly than alternatives. (A
leased telephone circuit, for example, represents a costly monthly
expense.) When line-of-sight transmission is possible, a microwave
link is a one-time expense that can offer greater bandwidth than a
leased circuit.
TRANSMISSION ME D IA
139
08.919-x CH03.i
140
8/28/98 12:02 PM
C h a p te r 3
Page 140
TR A NS M I S S I O N MEDIA
Costs are on the way down for low-power microwave systems for the
office. Although these systems don’t compete directly in cost with
cabled networks, microwave can be a cost-effective technology when
equipment must be moved frequently. Capacity can be extremely
high, but most data communication systems operate at data rates
between 1 and 10Mbps. Attenuation characteristics are determined
by transmitter power, frequency, and antenna size. Properly designed
systems are not affected by attenuation under normal operational
conditions; rain and fog, however, can cause attenuation of higher
frequencies.
Microwave systems are highly susceptible to atmospheric interference
and also can be vulnerable to electronic eavesdropping. For this reason, signals transmitted through microwave are frequently encrypted.
Satellite Microwave
Satellite microwave systems relay transmissions through communication satellites that operate in geosynchronous orbits 22,300 miles
above the earth. Satellites orbiting at this distance remain located
above a fixed point on earth.
Earth stations use parabolic antennas (satellite dishes) to communicate with satellites. These satellites then can retransmit signals in
broad or narrow beams, depending on the locations set to receive the
signals. When the destination is on the opposite side of the earth, for
example, the first satellite cannot transmit directly to the receiver
and thus must relay the signal through another satellite.
Because no cables are required, satellite microwave communication is
possible with most remote sites and with mobile devices, which
enables communication with ships at sea and motor vehicles.
The distances involved in satellite communication result in an interesting phenomenon: Because all signals must travel 22,300 miles to
the satellite and 22,300 miles when returning to a receiver, the time
required to transmit a signal is independent of distance on the
ground. It takes as long to transmit a signal to a receiver in the same
state as it does to a receiver a third of the way around the world. The
time required for a signal to arrive at its destination is called propagation delay. The delays encountered with satellite transmissions
range from 0.5 to 5 seconds.
Unfortunately, satellite communication is extremely expensive.
Building and launching a satellite can cost easily in excess of a billion
dollars. In most cases, organizations share these costs or purchase services from a commercial provider. AT&T, Hughes Network Services,
08.919-x CH03.i
8/28/98 12:02 PM
Page 141
C hapter 3
TRANSMISSION ME D IA
141
and Scientific-Atlanta are among the firms that sell satellite-based
communication services.
$Isatellite microwave
transmission>
Satellite links operate in the low gigahertz range, typically at
11–14GHz. Costs are extremely high and usually are distributed
across many users when communication services are sold. Bandwidth
is related to cost, and firms can purchase almost any required bandwidth. Typical data rates are 1–10Mbps. Attenuation characteristics
depend on frequency, power, and atmospheric conditions. Properly
designed systems also take attenuation into account. (Rain and
atmospheric conditions might attenuate higher frequencies.)
Microwave signals also are sensitive to EMI and electronic eavesdropping, so signals transmitted through satellite microwave frequently are encrypted as well.
Earth stations can be installed by numerous commercial providers.
Transmitters operate on licensed frequencies and require an FCC
license.
Comparisons of Different Wireless
Media
R E V I E W
B R E A K
The summary table below compares the different types of Wireless
communication media in terms of cost, ease of installation, distance
and “other issues.”
TABLE 3.3
C O M PA R I S O N
OF
WIRELESS MEDIA
Cable Type
Cost
Infrared
Cheapest of all the
wireless
Laser
Installation
Distance
Other Issues
Fairly easy, may require
line-of-sight
Under a kilometer
Can attenuate due to fog and rain
Similar to infrared
Requires line-of-sight
Can span several
kilometers
Can attenuate due to fog and rain
Narrow-band
radio
More expensive than
infrared and laser;
may need FCC license
Requires trained
technicians and can
involve tall radio towers
Can span hundreds
of kilometers
Low-power devices can attenuate; can
be eavesdropped upon; can also attenuate
due to fog, rain, and solar flares
Spread-spectrum
radio
More advanced technology
than narrow band radio,
thus more expensive
Requires trained
technicians and can
involve tall radio towers
Can span hundreds
of kilometers
Low-power devices can attenuate;
Can also attenuate due to fog, rain, and
solar flares
Microwave
Very expensive, as
requires link up to
satellites often
Requires trained
technicians and can
involve satellite dishes
Can span thousands
of kilometers
Can be eavesdropped upon;
can also attenuate due to fog,
rain, and solar flares
08.919-x CH03.i
142
8/28/98 12:02 PM
C h a p te r 3
Page 142
TR A NS M I S S I O N MEDIA
C A S E S T U DY : A N A LY Z I N G T R A N S M I S S I O N M E D I A N E E D S
ESSENCE OF THE CASE
The essential facts and features of this case
are as follows:
• Cost is an issue in Part 1.
• Cost and distance are an issue in Part 2.
• Security of data and speed is an issue in
Part 3.
• Traveling remote locations are the issue
in Part 4.
SCENARIO
The purpose of this case study is to put this
entire chapter into perspective. You saw from the
previous two chapters that a network is a connected set of devices. These can be computers,
printers, and servers, to name just a few of the
possible devices. These devices are networked
so that users can utilize different services on the
network. These services might be file and print
services, databases, or communication services.
To connect all these services together, some
form of transmission media must exist between
the devices on the network. As seen in Chapter
2, “Networking Standards,” the transmission
media operate at the Physical layer of the OSI
model. This chapter presented many forms of
transmission media.
To apply your knowledge of transmission media,
analyze the following evolving company. Notice
how the company’s business evolution leads to
different transmission media selections, regardless of the services used by the company.
Remember, whether a company is trying to give
file and print access to its users or access to a
database, some form of transmission media is
needed to connect the users of the services to
the services themselves. The case study is divided into four parts, each part representing a
growth stage of the company. The company in
question is called Mining Enterprises, and does
geological surveying.
Part 1
To begin with, Mining Enterprises is a small company with fifteen employees. They have just
opened shop in a small office complex. They
need to install a LAN, because they have an
informational database that all the employees
use, for purposes of payroll, accounting, and for
the geological informational database. Because
money is fairly tight, the company decides to
spend as little as possible to set up its network.
Part 2
Now, two years after installing its first LAN,
Enterprise Mining needs to expand. Business
has been very good, and employees are extremely productive working on an efficient LAN. The
problem is, though, that there is no office space
left for Enterprise Mining on its present floor, so
it needs to expand onto the 22nd floor. (It is currently on the 2nd floor.) Enterprise Mining needs
to connect its LAN on the second floor with the
LAN on the 22nd floor. Although business is
good, Enterprise Mining is still a little tight for
cash.
Part 3
It is now five years later. Enterprise Mining has
expanded even further. It now operates on eight
different floors. Each floor is almost like its own
08.919-x CH03.i
8/28/98 12:02 PM
Page 143
C hapter 3
TRANSMISSION ME D IA
143
C A S E S T U DY : A N A LY Z I N G T R A N S M I S S I O N M E D I A N E E D S
business unit, but a fair bit of data is still transferred between the different floors. Also, some
industrial espionage rumors have begun to circulate, so security is of importance. The budget
can be sacrificed to a degree for security, but the
sky is not the limit.
Part 4
The company has expanded into washing carpets
as well (nothing like a diversified company). They
now have a fleet of trucks that roam around
town, downloading information between the head
office and the trucks. The carpet cleaning business is very competitive, and Enterprise Mining
does not want the competition to be able to
intercept any information.
A N A LY S I S
Part 1
No requirements are mentioned that necessitate
the use of wireless media. Because costs are
the main concern, the possible bounded transmission media choices available are UTP, STP,
Fiber, or coaxial cable. The fiber cable is the
most expensive option, whereas UTP is the
cheapest. STP and coaxial fall somewhere
between. The media of choice for Part 1 is UTP,
unless they were encountering some form of EMI
that would require a transmission media that has
better shielding.
Part 2
Cost is still an issue, but so is distance. Two
solutions are possible. One is to go with the
cheapest cable type, but place a repeater on this
cable. This solution needs a cost estimate for
the price of cable and a repeater.
Another alternative is to move to a Thinnet or
Thicknet coax cable. The Thicknet cable costs
more than the Thinnet, and is probably not needed to span the 20 floors difference. This solution
involves only cable costs and no repeater costs.
The cost of laying the cable should be the same
in both cases. You would probably find that the
price of the Thinnet coax cable would be the
cheapest alternative in this case.
Part 3
Because data transfer between the eight business units is heavy, we probably would like to
use something with high bandwidth capability.
The decision would undoubtedly reflect a choice
to use a bound transmission media again. The
higher bandwidths are found in coaxial cable and
fiber-optic cable. Between these two options,
fiber-optic cable has a better resistance to eavesdropping. Because security is a concern, a
choice to use fiber optical cable is likely.
Part 4
This situation definitely leads to the use of some
form of wireless media. These vans probably are
moving around all the time and do not have a
line of sight with the head office. Due to the
movement, infrared and laser technologies
should be ruled out. Because the vans are probably going to be out of urban areas at times, this
rules out cellular media as well. This leaves
either a microwave solution or some type of radio
transmission.
continues
08.919-x CH03.i
144
8/28/98 12:02 PM
C h a p te r 3
Page 144
TR A NS M I S S I O N MEDIA
C A S E S T U DY : A N A LY Z I N G T R A N S M I S S I O N M E D I A N E E D S
continued
In analyzing microwave options, terrestrial
microwave could be an option, but this technology is used primarily to connect stationary sites.
Satellite microwave would probably be too costly
as an option.
Of the two remaining options (narrow-band radio
and spread-spectrum radio) spread-spectrum
radio offers a higher level of security. This is the
option most likely to be selected.
CHAPTER SUMMARY
KEY TERMS
Before taking the exam, make sure you
are familiar with the definitions of and
concepts behind each of the following
key terms. You can use the glossary
(Appendix A) for quick reference purposes.
• transmission media
• bounded media
• boundless media
• electromagnetic spectrum
• Electromagnetic Interference
(EMI)
• bandwidth
• attenuation
• baseband
This chapter examined the characteristics of some common network
transmission media. As explained in Chapter 2, transmission media
falls under the Physical layer of the OSI model. Regardless of what
services a network is providing, there must be some mechanism to
connect to these services.
This chapter provided some of the features of popular transmission
media. This chapter analyzed these features along the following
terms:
á Cost
á Ease of installation
á Distance limitation
á Bandwidth usage
á EMI characteristics
The major classifications of transmission media were broken down
into the following categories:
á Cable Media
• broadband
• UTP
• multiplexing
• STP
• frequency-division multiplexing
• Coaxial Cable
• time-division multiplexing
• Fiber Optic
08.919-x CH03.i
8/28/98 12:02 PM
Page 145
C hapter 3
TRANSMISSION ME D IA
145
CHAPTER SUMMARY
á Wireless Media
• Infrared
• Laser
• Narrow-band radio
• Spread-spectrum radio
• Microwave
• coaxial cable
• Thinnet
• Thicknet
• T-connector
• vampire clamp
• twisted-pair cable
Each form of transmission media was analyzed and compared in
terms of each evaluation criteria. The purpose of this chapter was
not to show which transmission media is best, but how each form
of transmission media had a unique set of characteristics that made
it adaptable to different situations and different sets of evaluation
criteria.
• unshielded twisted-pair cable
(UTP)
Cable media are often cheaper than wireless media, yet cable media
are also limited in the distances they can cover. Wireless media are
often more susceptible to EMI than fiber-optic cable is, but wireless
media are not subject to the accessibility and other installation problems faced by cable. In conclusion, each transmission media should
be evaluated in terms of the obstacles one will face in trying to relay
a signal from one device on the network to another.
• wireless media
• shielded twisted-pair cable (STP)
• fiber-optic cable
• IBM cabling
• infrared transmissions
• laser transmissions
• narrow-band radio
• spread-spectrum radio
• terrestrial microwave
• satellite microwave
08.919-x CH03.i
146
8/28/98 12:02 PM
C h a p te r 3
Page 146
TR A NS M I S S I O N MEDIA
A P P LY Y O U R L E A R N I N G
Exercises
3.1
Possible Solutions:
Sites that are far apart, that do not have the right
to lay cable between their buildings, need to
select some form of wireless media. Possible solutions that do not require a line of sight are
Choosing Transmission Media
Objective: To explore the possibilities of different
transmission media being used for different network
setups.
• Narrow-band radio transmission
Estimated time: 25 minutes
• Spread-spectrum radio transmission
This chapter presented a wide range of transmission
media possibilities. The purpose of this exercise is to
explore situations where different transmission media
could be used.
• Satellite microwave
1. Company A wants to set up a LAN. There is
EMI present in the building. What choices are
available to this company? What may be the
cheapest solution for this company?
Possible Solutions:
Almost all LANs use some form of bound media.
The five main choices in terms of cheapest to
most expensive are UTP, Thinnet, STP, Thicknet,
and fiber-optic. To actually solve this question,
one would need to test the degree of EMI interference. After the magnitude of this EMI is established, you can reduce the number of the cable
types that are possibilities. For example, if the
EMI was such that only Thicknet and fiber-optic
cable were feasible options, you would probably
select Thicknet to be your cable of choice,
because it is the cheapest solution of the two.
2. Company B wants to connect two sites together.
These sights are miles apart, with no line of sight
between the two buildings. The company has no
access rights on the land between the buildings.
What transmission media would be available to
them?
3.2
Shopping for Network Cabling
Objective: Explore the prices and availability of network cabling media in your area. Obtain a real-world
view of cabling options.
Estimated time: 15 minutes
This chapter discussed the advantages and disadvantages of common network transmission media. In this
exercise, you’ll explore how network installation professionals perceive the differences between the cabling
types. Remember that the cabling types discussed in
this chapter are all tied to particular network topologies
and architectures. You may want to read through
Chapter 4, “Network Topologies and Architectures,”
before attempting this exercise.
1. Call a local computer store (preferably a store
that provides network installations) and ask for
some basic information on network cabling. Ask
about coaxial Thinnet and Thicknet, UTP, and
STP. Learn with which type the store prefers to
work and in what situations they would recommend each of the types. Ask for pricing on
Thinnet PVC and plenum-grade cable. Try to get
a feeling for how the real world perceives the
cabling types described in this chapter.
08.919-x CH03.i
8/28/98 12:02 PM
Page 147
C hapter 3
TRANSMISSION ME D IA
147
A P P LY Y O U R L E A R N I N G
2. Computer vendors generally are busy people, so
try to be precise. Don’t imply that you’re getting
ready to buy a whole network (unless you are).
Just tell them you’re trying to learn more about
network cabling—vendors are often happy to
share their knowledge. If they’re helpful, remember them the next time you need a bid.
C. Witch widgets
D. Skeleton clamps
3. Which two of the following are true about UTP?
A. You can use an RJ-11 connector with an
RJ-45 socket.
B. UTP has the highest cost of any cabling system except Thinnet.
Review Questions
1. What are the two types of twisted pair media?
2. What are the names of two common types of
coaxial cable?
3. What is a major benefit of fiber-optic cable?
What is a major drawback of fiber-optic cable?
4. What are some reasons a wireless media would be
chosen over a bound media?
C. Telephone systems use UTP.
D. UTP is more sensitive to EMI than Thinnet.
4. Which of the following is not a permissible location for coaxial PVC cabling?
A. A bathroom
B. Above a drop ceiling
C. Outside
D. Along an exterior wall
Exam Questions
1. Which two of the following are true about coaxial Thinnet?
5. UTP Category 3 uses how many twisted pair(s)
of cables?
A. 1
B. 2
A. Thinnet cable is approximately 0.5 inches
thick.
C. 4
B. Thinnet has 50-ohm impedance.
D. 8
C. Thinnet is sometimes called Standard
Ethernet.
D. Thinnet cable includes an insulation layer.
2. Transceivers for Thicknet cables are often connected using what device?
A. Ghost taps
B. Vampire taps
6. Transmission rates of what speed are typical for
fiber-optic cables?
A. 10Mbps
B. 25Mbps
C. 100Mbps
D. 500Mbps
08.919-x CH03.i
148
8/28/98 12:02 PM
C h a p te r 3
Page 148
TR A NS M I S S I O N MEDIA
A P P LY Y O U R L E A R N I N G
7. What is a transceiver that connects a wireless
node with the LAN?
A. An access provider
B. Easy installation
C. High resistance to EMI due to twists in cable
D. Cabling of up to 500 meters
B. An access point
C. A Central Access Device (CAD)
D. A Wireless Access Device (WAD)
12. What are two benefits of shielding in a cable?
A. Reduction in signal attenuation
B. Reduction in EMI radiation
8. What type of transmissions are designed to reflect
the light beam off walls, floors, and ceilings until
it finally reaches the receiver?
A. Reflective infrared
B. Scatter infrared
C. Reduction in sensitivity to outside interference
D. None of the above
13. What are two disadvantages of fiber-optic cable?
C. Spread-spectrum infrared
A. Sensitive to EMI
D. None of the above
B. Expensive hardware
9. Which three of the following are forms of mobile
network technology?
A. Cellular
B. Packet-radio
C. Expensive to install
D. Limited in bandwidth
14. Which cable type is ideal for connecting between
two buildings?
C. UTP
A. UTP
D. Satellite station
B. STP
10. Which of the following cable types supports the
greatest cable lengths?
A. Unshielded twisted-pair
B. Shielded twisted-pair
C. Coaxial
D. Fiber-optic
15. What do radio transmissions require more of as
frequency increases?
C. Thicknet coaxial cable
Increasingly ______.
D. Thinnet coaxial cable
A. Attenuated
11. What are two advantages of UTP cable?
A. Low cost
B. Rapid
C. Line-of-sight
D. Sensitive to electromagnetic interference
08.919-x CH03.i
8/28/98 12:02 PM
Page 149
C hapter 3
TRANSMISSION ME D IA
149
A P P LY Y O U R L E A R N I N G
16. Which two statements are true of microwave systems?
20. Which two statements are true of Thinnet
cabling?
A. Microwave transmissions do not attenuate
under any conditions.
A. A T-connector must be used to connect the
PC’s network board to the network.
B. All microwave systems operate in the lowgigahertz range.
B. Either end of the cable can be terminated, but
not both ends.
C. Microwave signals are sensitive to EMI and
electronic eavesdropping.
C. BNC connectors cannot be used.
D. Unlike most other types of radio transmitters,
microwave transmitters don’t need to be
licensed.
17. For what are DIN Connectors primarily used?
A. Connecting UTP cables
B. Cabling Macintosh computers to AppleTalk
networks
C. Connecting devices with Thick-wire Ethernet
D. None of the above
18. Which two connectors are frequently used with
STP cable?
A. T-connectors
B. RJ-45 connectors
C. IBM unisex connectors
D. AppleTalk DIN connectors
19. Which two connectors are commonly used with
coaxial cable?
A. DB-25 connectors
B. T-connectors
C. ST-connectors
D. BNC connectors
D. One terminator must be grounded.
21. Which form of spread-spectrum media breaks
data into chips, which are transmitted on separate
frequencies?
A. Frequency hopping
B. Data spread
C. Frequency circulation
D. Direct sequence modulation
22. What wireless system typically operates in the low
gigahertz range?
A. Laser
B. Terrestrial microwave
C. Infrared
D. Audible sound
23. What is the term used to describe the time
required for a signal to arrive at its destination in
a satellite microwave system?
A. Propagation delay
B. Modulation delay
C. Transmit delay
D. Session delay
08.919-x CH03.i
150
8/28/98 12:02 PM
C h a p te r 3
Page 150
TR A NS M I S S I O N MEDIA
A P P LY Y O U R L E A R N I N G
24. You are to choose a transmission media type for a
network. The capacity for intruders to “sniff ”
information from the network is a major concern.
Also EMI is a major consideration.
Primary Objective: The transmission media must
be capable of transferring the data over ten miles.
Secondary Objective: Electrical lightning storms
are common in the area, so the transmission
media needs to be independent of the weather.
Secondary Objective: The transmission media
needs to be relatively inexpensive.
Suggested Solution: Implement the network
using fiber-optic cabling.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objectives.
D. This solution does not satisfy the primary
objective.
The major drawback with fiber-optic cable is its
cost.
4. Some typical situations that call for wireless
media are
• Spaces where cabling would be impossible or
inconvenient. These include open lobbies,
inaccessible parts of buildings, older buildings, historical buildings where renovation is
prohibited, and outdoor installations.
• People who move around a lot within their
work environments. Network administrators,
for instance, must troubleshoot a large office
network. Nurses and doctors need to make
rounds at a hospital.
• Temporary installations. These situations
include any temporary department set up for
a specific purpose that soon will be torn down
or relocated.
• People who travel outside the work environment and need instantaneous access to network resources.
• Satellite offices or branches that need to be
connected to a main office or location.
Answers to Review Questions
1. The two major types of twisted pair cabling are
shielded twisted-pair (STP) and unshielded
twisted-pair (UTP). STP has better EMI protection.
2. The two most common types of coax cable are
Thinnet and Thicknet.
3. The major benefits of fiber-optic cable are immunity to EMI, high bandwidth, and the long distances that a cable can run.
Answers to Exam Questions
1. B, D. Thinnet cable includes an insulation layer
and needs a 50-ohm terminator. See “Thinnet”
under “Cable Media” section in this chapter.
2. B. A vampire clamp is used to clamp a transceiver
onto a Thicknet cable. See “Thicknet” under
“Cable Media” section in this chapter.
08.919-x CH03.i
8/28/98 12:02 PM
Page 151
C hapter 3
TRANSMISSION ME D IA
151
A P P LY Y O U R L E A R N I N G
3. C, D. Telephone cable is UTP, and UTP has the
highest sensitivity to EMI. See “Unshielded
Twisted-Pair (UTP) Cable” under “Cable Media.”
14. D. Fiber is the preferred medium between buildings when using a bound transmission media. See
“Fiber-Optic Cable” under “Cable Media.”
4. B. PVC emits toxic fumes when it burns and is
not permitted in plenum spaces. See “Coax and
Fire Code Classifications” under “Cable Media.”
15. C. The higher the frequency, the more of a line of
sight is required. See “Wireless Communications
with LANs” under “Wireless Media.”
5. C. Category 3 UTP uses 4 pairs of twisted-pair
cables. See “Unshielded Twisted-Pair (UTP)
Cable” under “Cable Media.”
16. B, C. A is simply false, and D is incorrect because
microwave transmissions do need to be licensed.
See “Wireless Media.”
6. C. Standard transmission rates for fiber-optic
cable are 100Mbps. See “Fiber-Optic Cable”
under “Cable Media.”
17. B. A DIN is used by Macintosh computers. See
“Connectors for STP” under “Cable Media.”
7. B. The function of an access point is to relay
information between a transceiver and the LAN.
See “Wireless Communications with LANs”
under “Wireless Media.”
8. B. Scatter infrared does not require line of sight.
See “Infrared Transmission” under “Wireless
Media.”
9. A, B, D. UTP is not a wireless technology.
Compare the sections titled “Cable Media” and
“Wireless Media.”
10. C. Thicknet supports the greatest lengths of all
the cable types listed. See “Thicknet” under
“Cable Media.”
11. A, B. C has to do with crosstalk; D is not true.
See “Unshielded Twisted-Pair (UTP) Cable”
under “Cable Media.”
12. B, C. B and C are why shielding is used. See the
section “Cable Media.”
13. B, C. A and D are not a factor with fiber-optic
cable. See “Fiber-Optic Cable” under “Cable
Media.”
18. C, D. A is for coaxial cables, whereas B is used
primarily with UTP. See the section “Cable
Media.”
19. B, D. T connectors attach to the BNC connector.
See “Coaxial Cable” under “Cable Media.”
20. A, D. Both ends need to be terminated, hence B
is incorrect. BNC connectors are used, hence C is
incorrect. See “Coaxial Cable” under “Cable
Media.”
21. D. Spread-spectrum media uses frequency hopping in general. The data is spread across different frequencies when being transmitted. This
spread of the data is circulated between the different frequencies being used. The actual term to
describe this is direct-sequence modulation. See
“Spread-Spectrum Radio Transmission” under
“Wireless Media.”
22. B. All other answers operate at a lower frequency
range. See “Microwave” under “Wireless Media.”
23. A. “Propagation delay” is the term used to explain
the delay that occurs when data is transmitted
within a satellite microwave system. This delay
causes a delay of a session being established and
08.919-x CH03.i
152
8/28/98 12:02 PM
C h a p te r 3
Page 152
TR A NS M I S S I O N MEDIA
A P P LY Y O U R L E A R N I N G
of data being transmitted. See “Satellite
Microwave” under “Wireless Media.”
24. B. Fiber-optic cable enables the network to span
many miles as well as be immune to weather
conditions. The last secondary objective will not
be met, because fiber optic cable solutions are
among the most expensive solutions to implement on the market.
Suggested Readings and Resources
1. Kayata Wesel, Ellen. Wireless Multimedia
Communications: Networking Video, Voice, and
Data. Addison-Wesley, 1997.
2. Horak, Ray, Uyless Black, and Mark Miller.
Communication Systems and Networks: Voice,
Data, and Broadband Technologies. IDG Books,
1996.
3. Black, Uyless. Computer Networks: Protocols,
Standards, and Interfaces—The Professional’s
Guide. Prentice-Hall, 1993.
09.919-x CH04.i
8/28/98 12:24 PM
Page 153
OBJECTIVES
Chapter 4 targets one multi-part objective in the
Planning section of the Networking Essentials exam:
Select the appropriate topology for various tokenring and ethernet networks.
. This objective is necessary because token-ring or
ethernet networks can utilize different physical and
logical topologies. This exam objective points out
the need for you to be able to identify which topology should be used by a token-ring or ethernet
network given different circumstances or environmental conditions.
C H A P T E R
4
Network Topologies
and Architectures
09.919-x CH04.i
8/28/98 12:24 PM
Page 154
OUTLINE
S T U DY S T R AT E G I E S
Access Methods
155
Contention
156
Polling
158
Token Passing
158
Comparing Contention and Token
Passing
160
Demand Priority
162
Network Topologies
162
Bus Topologies
163
Ring Topologies
164
Star Topologies
164
Mesh Topology
165
Network Architectures
166
Ethernet
Ethernet Cabling
167
169
Token Ring
Token Ring Cabling
Passing Data on Token Rings
The Beaconing Process
178
179
181
183
ARCNet
183
FDDI
185
Chapter Summary
188
. This chapter addresses physical and logical
topology types. You can deploy a range of
different physical and logical topologies on your
network. Token-ring and ethernet networks can
utilize some, but not necessarily all, of these
different physical and logical topologies. Be
aware of the advantages and disadvantages of
the different topologies and which ones can be
used by and with token-ring or ethernet networks.
09.919-x CH04.i
8/28/98 12:24 PM
Page 155
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
INTRODUCTION
Networks come in a few standard forms or architectures, and each
form is a complete system of compatible hardware, protocols, transmission media, and topologies. A topology is a map of the network.
It is a plan for how the cabling will interconnect the nodes, or
devices, and how the nodes will function in relation to one another.
Several factors shape the various network topologies, and one of the
most important is the choice of an access method. An access method
is a set of rules for sharing the transmission medium. This chapter
describes two of the most important categories of access methods:
contention and token passing. You learn about CSMA/CD and
CSMA/CA, two contention-based access methods, and about some
of the fundamental topology archetypes. This chapter then looks at
ethernet, token-ring, ARCNet, and FDDI networks. These types of
networks all utilize either a contention-based or token-passing access
method.
The exam objective being addressed focuses on a selection of the
appropriate topology for either ethernet or token-ring architectures.
As you read this chapter be very aware of the differences between a
physical and logical topology, because these terms mean different
things. Pay particular attention to the cabling specifications used by
the different topologies. ARCNet and FDDI are discussed in this
chapter so that the subject of network topologies and architectures
can be addressed completely.
ACCESS METHODS
An access method is a set of rules governing how the network nodes
share the transmission medium. The rules for sharing among computers are similar to the rules for sharing among humans in that
they both boil down to a pair of fundamental philosophies:
1) first come, first served and 2) take turns. These philosophies are the
principles defining the three most important types of media access
methods:
á Contention. In its purest form, contention means that the com-
puters are contending for use of the transmission medium.
Any computer in the network can transmit at any time (first
come, first served).
155
09.919-x CH04.i
156
8/28/98 12:24 PM
C h a p te r 4
Page 156
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
á Polling. One device is responsible for polling the other devices
to see whether they are ready for the transmission or reception
of data.
á Token passing. The computers take turns using the transmission
medium.
As you can imagine, contention-based access methods can give rise
to situations in which two or more of the network nodes try to
broadcast at the same time and the signals collide. Specifications for
contention-based access methods include procedures for how to
avoid collisions and what to do if a collision occurs. This section
introduces the CSMA/CD and CSMA/CA access methods.
On most contention-based networks, the nodes are basically equal.
No node has a higher priority than other nodes. A new access
method called demand priority, however, resolves contention and collisions and in so doing accounts for data type priorities. This section
also describes demand priority access.
Contention
In pure contention-based access control, any computer can transmit
at any time. This system breaks down when two computers attempt
to transmit at the same time, in which case a collision occurs (see
Figure 4.1). Eventually, when a network gets busy enough, most
attempts to transmit result in collisions and little effective communication can take place.
Mechanisms are usually put into place to minimize the number of
collisions. One mechanism is carrier sensing, whereby each computer
listens to the network before attempting to transmit. If the network
is busy, the computer refrains from transmitting until the network
quiets down. This simple “listen before talking” strategy can significantly reduce collisions.
Another mechanism is carrier detection. With this strategy, computers continue to listen to the network as they transmit. If a computer
detects another signal that interferes with the signal it’s sending, it
stops transmitting. Both computers then wait a random amount of
time and attempt to retransmit. Unless the network is extremely
busy, carrier detection along with carrier sensing can manage a large
volume of transmissions.
09.919-x CH04.i
8/28/98 12:24 PM
Page 157
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
collision
Carrier detection and carrier sensing used together form the protocol used in all types of ethernet: Carrier Sense Multiple Access with
Collision Detection (CSMA/CD). CSMA/CD limits the size of the
network to 2,500 meters. At longer distances, the broadcast-sensing
mechanisms don’t work—a node at one end can’t sense when a node
at the other end starts to broadcast.
Apple’s LocalTalk (See Chapter 7, “Transport Protocols,” for more
details) network uses the protocol Carrier Sense Multiple Access with
Collision Avoidance (CSMA/CA). Collision avoidance uses additional
techniques to further reduce the likelihood of collisions. In
CSMA/CA, each computer signals a warning that says it is about to
transmit data, and then the other computers wait for the broadcast.
CSMA/CA adds an extra layer of order, thereby reducing collisions,
but the warning broadcasts increase network traffic, and the task of
constantly listening for warnings increases system load.
CSMA/CD can be compared to trying to walk across the street and
almost being hit by a car. If you are almost hit by a car, then you
wait a few moments before trying to cross again. CSMA/CA is similar, but in this case you send your friend across the street first. If
your friend is almost hit by a car, then you wait. If a car does not hit
him, then you proceed.
Although it sounds as if contention methods are unworkable due to
the risk of collisions, contention (in particular CSMA/CD in the
form of ethernet) is the most popular media access control method
on LANs. In fact, no currently employed LAN standards utilize pure
contention access control without adding some mechanism to
reduce the incidence of collisions.
FIGURE 4.1
A collision on a contention-based network.
157
09.919-x CH04.i
158
8/28/98 12:24 PM
C h a p te r 4
Page 158
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
Contention is a simple protocol that can operate with simple network software and hardware. Unless traffic levels exceed about 30%
of bandwidth, contention works quite well. Contention-based networks offer good performance at low cost.
Because collisions occur at unpredictable intervals, no computer is
guaranteed the capability to transmit at any given time. Contentionbased networks are called probabilistic because a computer’s chance
of being permitted to transmit cannot be precisely predicted.
Collisions increase in frequency as more computers use the network.
When too many computers use the network, collisions can dominate
network traffic, and few frames are transmitted without error.
All computers on a contention-based network are equal.
Consequently, it’s impossible to assign certain computers higher priorities and, therefore, greater access to the network.
Contention access control is well-suited for networks that experience
bursts in traffic (such as large intermittent file transfers, for instance)
and have relatively few computers.
Polling
Polling-based systems require a device (called a controller, or master
device) to poll other devices on the network to see whether they are
ready to either transmit or receive data as seen in Figure 4.2. This
access method is not widely used on networks because the polling
itself can cause a fair amount of network traffic. A common example
of polling is when your computer polls its printer to receive a print
job.
Token Passing
Token passing utilizes a frame called a token, which circulates around
the network. A computer that needs to transmit must wait until it
receives the token, at which time the computer is permitted to transmit. When the computer is done transmitting, it passes the token
frame to the next station on the network. Figure 4.3 shows how
token passing is implemented on a token-ring network. Token-ring
networks are discussed in greater detail later in this chapter in the
section titled “Token Ring.”
09.919-x CH04.i
8/28/98 12:25 PM
Page 159
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
FIGURE 4.2
A
Ready to transmit node B?
An example of polling-based access.
B
C
FIGURE 4.3
Token passing.
R
T
T
R
R
T
TOKEN
T
R
159
09.919-x CH04.i
160
8/28/98 12:25 PM
C h a p te r 4
Page 160
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
Several network standards employ token passing access control:
á Token ring. The most common token-passing standard,
embodied in IEEE standard 802.5.
á IEEE standard 802.4. Implemented infrequently; defines a bus
network that also employs token passing. ARCNet can deploy
this standard, as is shown later in the chapter in the section
titled “ARCNet.”
á FDDI. A 100Mbps fiber-optic network standard that uses
token passing and rings in much the same manner as 802.5
token ring.
Token-passing methods can use station priorities and other methods
to prevent any one station from monopolizing the network. Because
each computer has a chance to transmit each time the token travels
around the network, each station is guaranteed a chance to transmit
at some minimum time interval.
Token passing is more appropriate than contention under the following conditions:
á When the network is carrying time-critical data. Because token
passing results in more predictable delivery, token passing is
called deterministic.
á When the network experiences heavy utilization. Performance
typically falls off more gracefully with a token-passing network
than with a contention-based network. Token-passing networks cannot become gridlocked due to excessive numbers of
collisions.
á When some stations should have higher priority than others. Some
token-passing schemes support priority assignments.
Comparing Contention and Token
Passing
As an access control mechanism, token passing appears to be clearly
superior to contention. You’ll find, however, that ethernet, by far the
dominant LAN standard, has achieved its prominence while firmly
wedded to contention access control.
Page 161
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
Token passing requires a variety of complex control mechanisms to
work well. The necessary hardware is considerably more expensive
than the hardware required to implement the much simpler contention mechanisms. The higher cost of token-passing networks is
difficult to justify unless the special features are required.
Token-Passing Throughput Figure
4.4 implies that token-passing
throughput eventually reaches a zero
level, but that cannot happen, regardless of the loading conditions.
Although a station’s access to the
network might be limited, the workstation is guaranteed the right to add
itself to the reservation list on each
circuit. It may take several more circuits of the token before the station’s
data is actually sent, but it will be
sent.
Hubs A hub is a device on the network that connects many short cables
together. It is discussed in detail in
Chapter 6, “Connectivity Devices and
Transfer Mechanisms.”
Because token-passing networks are designed for high reliability,
building network diagnostic and troubleshooting capabilities into
the network hardware is common. These capabilities increase the
cost of token-passing networks. Organizations must decide whether
this additional reliability is worth the extra cost.
Conversely, although token-passing networks perform better than
contention-based networks when traffic levels are high, contention
networks exhibit superior performance under lighter load conditions. Passing the token around (and other maintenance operations)
eats into the available bandwidth. As a result, 10Mbps ethernet and
16Mbps token-ring networks perform comparably well under light
load conditions, but the ethernet costs considerably less.
THROUGHPUT
Figure 4.4 illustrates the performance characteristics you can expect
from each access control method.
161
NOTE
8/28/98 12:25 PM
NOTE
09.919-x CH04.i
Token passing
Contention
LOAD
FIGURE 4.4
Comparison of contention and token passing.
09.919-x CH04.i
162
8/28/98 12:25 PM
C h a p te r 4
Page 162
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
Demand Priority
Demand priority is an access method used with the new 100Mbps
100VG-AnyLAN standard. Although demand priority is officially
considered a contention-based access method, demand priority is
considerably different from the basic CSMA/CD ethernet. In
demand priority, network nodes are connected to hubs, and those
hubs are connected to other hubs. Contention, therefore, occurs at
the hub. (100VG-AnyLAN cables can actually send and receive data
at the same time.) Demand priority provides a mechanism for prioritizing data types. If contention occurs, data with a higher priority
takes precedence.
NETWORK TOPOLOGIES
A topology defines the arrangement of nodes, cables, and connectivity devices that make up the network. Two categories form the basis
for all discussions of topologies:
á Physical topology. Describes the actual layout of the network
transmission media.
á Logical topology. Describes the logical pathway a signal follows
as it passes among the network nodes.
Another way to think about this distinction is that a physical topology defines the way the network looks, and a logical topology defines
the way the data passes among the nodes. At a glance this distinction
may seem nit-picky, but as you will learn in this chapter, the physical
and logical topologies for a network can be very different. A network
with a star physical topology, for example, may actually have a bus
or a ring logical topology.
In common usage, the word “topology” applies to a complete network definition, which includes the physical and logical topologies
and also the specifications for elements such as the transmission
medium. The term topology as used in Microsoft’s test objectives for
the Networking Essentials exam is not limited to the physical and
logical topology archetypes (that is, the design or layout) described
in this section. It applies to the complete network specifications
(such as 10BASE-T or 10BASE5) described in the “Ethernet” and
“Token Ring” sections of this chapter.
09.919-x CH04.i
8/28/98 12:25 PM
Page 163
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
Physical and logical topologies can take several forms. The most
common and the most important for understanding the ethernet
and token-ring topologies (described later in this chapter) are the
following:
á Bus topologies
á Ring topologies
á Star topologies
á Mesh topology
The following sections discuss each of these important topology
types.
Bus Topologies
A bus physical topology is one in which all devices connect to a common, shared cable (sometimes called the backbone). A bus physical
topology is shown in Figure 4.5.
If you think the bus topology seems ideally suited for the networks
that use contention-based access methods such as CSMA/CD,
you are correct. Ethernet, the most common contention-based network architecture, typically uses bus as a physical topology. Even
10BASE-T ethernet networks (described later in this chapter) use
the bus as a logical topology but are configured in a star physical
topology.
FIGURE 4.5
A bus physical topology.
163
09.919-x CH04.i
164
8/28/98 12:25 PM
C h a p te r 4
Page 164
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
Most bus networks broadcast signals in both directions on the backbone cable, enabling all devices to directly receive the signal. Some
buses, however, are unidirectional: Signals travel in only one direction and can reach only downstream devices. Recall from Chapter 3,
“Transmission Media,” that a special connector called a terminator
must be placed at the end of the backbone cable to prevent signals
from reflecting back on the cable and causing interference. In the
case of a unidirectional bus, the cable must be terminated in such a
way that signals can go down the cable but do not reflect back up
the cable and reach other devices, causing disruption.
T
Ring Topologies
R
R
T
T
R
R
T
T = TRANSMIT
R = RECEIVE
Ring topologies are ideally suited for token-passing access methods.
The token passes around the ring, and only the node that holds the
token can transmit data.
Ring physical topologies are quite rare. The ring topology is almost
always implemented as a logical topology. Token ring, for example,
the most widespread token-passing network, always arranges the
nodes in a physical star (with all nodes connecting to a central hub),
but passes data in a logical ring (see Figure 4.7).
A ring topology.
You get a closer look at token ring later in this chapter in the section
titled “Token Ring.”
NOTE
FIGURE 4.6
Ring topologies are wired in a circle. Each node is connected to its
neighbors on either side, and data passes around the ring in one
direction only (see Figure 4.6). Each device incorporates a receiver
and a transmitter and serves as a repeater that passes the signal on to
the next device in the ring. Because the signal is regenerated at each
device, signal degeneration is low.
Star Topologies
The Star Physical Topology A star
physical topology means that the
nodes are all connected to a central
hub. The path the data takes among
the nodes and through that hub (the
logical topology) depends on the
design of the hub, the design of the
cabling, and the hardware and software configuration of the nodes.
Star topologies require that all devices connect to a central hub (see
Figure 4.8). The hub receives signals from other network devices and
routes the signals to the proper destinations. Star hubs can be interconnected to form tree, or hierarchical, network topologies.
As mentioned earlier, a star physical topology is often used to implement a bus or ring logical topology (refer to Figure 4.5).
09.919-x CH04.i
8/28/98 12:26 PM
Page 165
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
165
FIGURE 4.7
A logical ring configuration in a physical star.
R
T
R
T
T
R
T
R
FIGURE 4.8
A star topology.
Mesh Topology
A popular test subject is the mesh topology. A mesh topology (see
Figure 4.9) is really a hybrid model representing an all-channel
sort of physical topology. It is a hybrid because a mesh topology
can incorporate all the topologies covered to this point. It is an
09.919-x CH04.i
166
8/28/98 12:26 PM
C h a p te r 4
Page 166
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
all-channel topology in that every device is directly connected to
every other device on the network. When a new device is added, a
connection to all existing devices must be made. This provides for a
great deal of fault tolerance, but it involves extra work on the part of
the network administrator. That is, if any transmission media breaks,
the data transfer can take alternative routes. However, cabling
becomes much more extensive and complicated.
FIGURE 4.9
PC
A mesh topology.
PC
PC
PC
These different connections can be the same (all ethernet) or different (a mix of ethernet and token ring).
NETWORK ARCHITECTURES
A network architecture is the design specification of the physical layout of connected devices. This includes the cable being used (or
wireless media being deployed), the types of network cards being
deployed, and the mechanism through which data is sent on to the
network and passed to each device. Network architecture, in short,
encompasses the total design and layout of the network.
09.919-x CH04.i
8/28/98 12:26 PM
Page 167
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
Ethernet
Ethernet is a very popular local area network architecture based on
the CSMA/CD access method. The original ethernet specification
was the basis for the IEEE 802.3 specifications (see Chapter 2,
“Networking Standards”). In present usage, the term “ethernet”
refers to original ethernet (or Ethernet II, the latest version) as well
as the IEEE 802.3 standards. The different varieties of ethernet
networks are commonly referred to as ethernet topologies. Typically,
ethernet networks can use a bus physical topology, although, as
mentioned earlier, many varieties of ethernet such as 10BASE-T use
a star physical topology and a bus logical topology. (Microsoft uses
the term “star bus topology” to describe 10BASE-T.)
Ethernet networks, depending on the specification, operate at 10- or
100Mbps using baseband transmission. Each IEEE 802.3 specification (see Chapter 2) prescribes its own cable types.
Later sections in this chapter examine the following ethernet
topologies:
á 10BASE2
á 10BASE5
á 10BASE-T
á 10BASE-FL
á 100VG-AnyLAN
á 100BASE-X
Note that the name of each ethernet topology begins with a number
(10 or 100). That number specifies the transmission speed for the
network. For instance, 10BASE5 is designed to operate at 10Mbps,
and 100BASE-X operates at 100Mbps. “BASE” specifies that baseband transmissions are being used. The “T” is for unshielded
twisted-pair wiring, “FL” is for fiber optic cable, “VG-AnyLAN”
implies Voice Grade, and “X” implies multiple media types.
Ethernet networks transmit data in small units called frames. The
size of an ethernet frame can be anywhere between 64 and 1,518
bytes. Eighteen bytes of the total frame size are taken up by frame
overhead, such as the source and destination addresses, protocol
167
09.919-x CH04.i
NOTE
168
8/28/98 12:26 PM
C h a p te r 4
Page 168
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
Ethernet Origins The origins of ethernet are commemorated in the initials DIX, a 15-pin connector used to
interface ethernet components. The
acronym “DIX” derives from the combination of leading letters of the
founding ethernet vendors: Digital,
Intel, and Xerox.
information, and error-checking information. There are many different types of ethernet frames, such as the Ethernet II, 802.2, and
802.3 frames to name a few. It is important to remember that 802.2
and 802.3 are IEEE specifications on how information is transferred
onto the transmission media (Data Link layer) as well as the specification on how the data should be packaged. More information on
frame types is discussed in Chapter 5, “Network Adapter Cards.”
A typical Ethernet II frame has the following sections:
á Preamble. A field that signifies the beginning of the frame.
á Addresses. A field that identifies the source and destination
addresses for the frame.
á Type. A field that designates the Network layer protocol.
á Data. The data being transmitted.
á CRC. Cyclical Redundancy Check for error checking.
These parts of the frame are illustrated in Figure 4.10.
The term “ethernet” commonly refers to original ethernet (which has
been updated to Ethernet II) as well as the IEEE 802.3 standards.
Ethernet and the 802.3 standards differ in ways significant enough
to make standards incompatible in terms of packet formats, however.
At the Physical layer, ethernet and 802.3 are generally compatible in
terms of cables, connectors, and electronic devices.
Ethernet generally is used on light-to-medium traffic networks and
performs best when a network’s data traffic transmits in short bursts.
Ethernet is the most commonly used network standard.
One advantage of the linear bus topology used by most ethernet networks (this doesn’t apply to star bus networks such as 10BASE-T) is
that the required cabling is minimized because a separate cable run
to the hub for each node is not required. One disadvantage is that a
break in the cable or a streaming network adapter card can bring
Preamble
Address
FIGURE 4.10
A sample of part of an Ethernet II frame.
Type
Data
CRC
09.919-x CH04.i
8/28/98 12:26 PM
Page 169
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
down the entire network. Streaming is more frequently referred to as
a broadcast storm. A broadcast storm occurs when a network card
fails and the transmitter floods the cable with traffic, like a faucet
stuck open. At this point, the network becomes unusable. See
Chapter 12, “Troubleshooting,” for more on broadcast storms.
Ethernet Cabling
You can use a variety of cables to implement ethernet networks.
Many of these cable types, such as Thinnet, Thicknet, UTP, and
STP, are described in Chapter 3. Ethernet networks traditionally
have used coaxial cables of several different types. Fiber-optic cables
now are frequently employed to extend the geographic range of ethernet networks.
The contemporary interest in using twisted-pair wiring has resulted
in a scheme for cabling that uses unshielded twisted-pair (UTP).
The 10BASE-T cabling standard uses UTP in a star physical topology. (10BASE-T is discussed later in this chapter.)
Ethernet remains closely associated with coaxial cable. Two types of
coaxial cable still used in small and large environments are Thinnet
(10BASE2) and Thicknet (10BASE5). Thinnet and Thicknet ethernet networks have different limitations that are based on the
Thinnet and Thicknet cable specifications. The best way to remember the requirements for ethernet cable types is to use the 5-4-3 rule
of thumb for each cable type.
The 5-4-3 rule (see Figure 4.11) states that the following can appear
between any two nodes in the ethernet network:
á Up to 5 segments in a series
á Up to 4 concentrators or repeaters
á 3 segments of cable (coaxial only) that contain nodes
The following subsections describe some of the characteristics of
cable types used in ethernet topologies.
10BASE2
The 10BASE2 cabling topology (Thinnet) generally uses the onboard transceiver of the network interface card to translate the signals to and from the rest of the network. Thinnet cabling, described
169
09.919-x CH04.i
170
8/28/98 12:27 PM
C h a p te r 4
Page 170
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
FIGURE 4.11
The 5-4-3 rule: 5 segments on a LAN, 4 connection devices (hubs or repeaters), and only 3
populated segments.
in Chapter 3, uses BNC T-connectors that attach directly to the network adapter. Each end of the cable should have a terminator, and
you must use a grounded terminator on one end (see Figure 4.12).
The main advantage of using 10BASE2 in your network is cost.
When any given cable segment on the network doesn’t have to be
run farther than 185 meters (607 feet), 10BASE2 is often the cheapest network cabling option.
T CONNECTOR
RG-58 CABLE
TO GROUND
TO OTHER
WORKSTATIONS
BNC CONNECTOR
FIGURE 4.12
T connector and a BNC connector.
TERMINATOR
09.919-x CH04.i
8/28/98 12:27 PM
Page 171
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
171
10BASE2 is also relatively simple to connect. Each network node
connects directly to the network cable with a T-connector attached
to the network adapter. For a successful installation, you must
adhere to several rules in 10BASE2 ethernet environments, including the following:
á The minimum cable distance between clients must be 0.5
meters (1.5 feet).
á Pig tails, also known as drop cables, from T-connectors
tion of 185 meters (607 feet).
á The entire network cabling scheme cannot exceed 925 meters
(3,035 feet).
á The maximum number of nodes per network segment is 30
(this includes clients and repeaters).
á A 50-ohm terminator must be used on each end of the bus
with only one of the terminators having either a grounding
strap or a grounding wire that attaches it to the screw holding
an electrical outlet cover in place.
á You may not have more than five segments on a network.
These segments may be connected with a maximum of four
repeaters, and only three of the five segments may have network nodes.
Figure 4.13 shows two network segments using 10BASE2 cabling.
For more on 10BASE2’s Thinnet cabling, see Chapter 3.
10BASE5
The 10BASE5 cabling topology (Thicknet) uses an external transceiver to attach to the network adapter card (see Figure 4.14). The
external transceiver clamps to the Thicknet cable (as described in
Chapter 3). An Attachment Universal Interface (AUI) cable runs
from the transceiver to a DIX connector on the back of the network
adapter card. As with Thinnet, each network segment must be terminated at both ends, with one end using a grounded terminator.
The components of a Thicknet network are shown in Figure 4.15.
EXAM
á You may not exceed the maximum network segment limita-
TIP
shouldn’t be used to connect to the BNC connector on the
network adapter. The T-connector must be connected directly
to the network adapter.
Metric Conversion You should be
able to translate cable segment
lengths from feet to meters or from
meters to feet. A meter is equivalent to 39.37 inches or 3.28 feet.
09.919-x CH04.i
172
8/28/98 12:27 PM
C h a p te r 4
Page 172
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
FIGURE 4.13
Two segments using 10BASE2 cabling.
Terminator
Transceiver
1.5 Feet Minimum
Repeater
Ground
607 Feet Maximum
FIGURE 4.14
Two segments using 10BASE5 cabling.
Ground
Wire
Transceiver
Cable
Terminator
with Ground
8 Feet Minimum
Transceiver
Repeater
Segment 2
1,640 Feet Maximum
09.919-x CH04.i
8/28/98 12:27 PM
Page 173
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
FIGURE 4.15
Thick Net Transceiver
Components of a Thicknet network.
N-Series Connector
Other
Workstations
N-Series
Barrel
Connector
DIX
Connector
N-Series
Terminator
Transceiver
Transceiver
Cable
The primary advantage of 10BASE5 is its capability to exceed the
cable restrictions that apply to 10BASE2. 10BASE5 does pose
restrictions of its own, however, which you should consider when
installing or troubleshooting a 10BASE5 network. As with 10BASE2
networks, the first consideration when you troubleshoot a 10BASE5
network should be the established cabling rules and guidelines. You
must follow several additional guidelines, along with the 5-4-3 rule,
when configuring Thicknet networks, such as the following:
á The minimum cable distance between transceivers is 2.5
meters (8 feet).
á You may not go beyond the maximum network segment
length of 500 meters (1,640 feet).
á The entire network cabling scheme cannot exceed 2,500
meters (8,200 feet).
á One end of the terminated network segment must be
grounded.
á Drop cables (transceiver cables) can be as short as required but
cannot be longer than 50 meters from transceiver to computer.
á The maximum number of nodes per network segment is 100.
(This includes all repeaters.)
173
09.919-x CH04.i
174
8/28/98 12:28 PM
C h a p te r 4
Page 174
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
The length of the drop cables (from the transceiver to the computer)
is not included in measurements of the network segment length and
total network length.
As Chapter 3 mentions, Thicknet and Thinnet networks are often
combined, with a Thicknet backbone merging smaller Thinnet segments. (See Chapter 3 for more on 10BASE5’s Thicknet cabling.)
10BASE-T
The trend in wiring ethernet networks is to use unshielded twistedpair (UTP) cable. 10BASE-T, which uses UTP cable, is also one of
the more popular implementations for ethernet. It is based on the
IEEE 802.3 standard. 10BASE-T supports a data rate of 10Mbps
using baseband.
10BASE-T cabling is wired in a star topology. The nodes are wired
to a central hub, which serves as a multiport repeater (see Figure
4.16). A 10BASE-T network functions logically as a linear bus. The
hub repeats the signal to all nodes, and the nodes contend for access
to the transmission medium as if they were connected along a linear
bus. The cable uses RJ-45 connectors, and the network adapter card
can have RJ-45 jacks built into the back of the card (An RJ-45 connector looks very similar to a telephone plug.)
TP Hub
FIGURE 4.16
A 10BASE-T network wired in a star topology.
Twisted-Pair Ethernet Cabling
09.919-x CH04.i
8/28/98 12:28 PM
Page 175
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
10BASE-T segments can be connected using coaxial or fiber-optic
backbone segments. Some hubs provide connectors for Thinnet and
Thicknet cables (in addition to 10BASE-T UTP-type connectors).
By attaching a 10BASE-T transceiver to the AUI port of the network adapter, you can use a computer set up for Thicknet on a
10BASE-T network.
The star wiring of 10BASE-T provides several advantages, particularly in larger networks. First, the network is more reliable and easier
to manage because 10BASE-T networks use a concentrator (a centralized wiring hub). These hubs are “intelligent” in that they can
detect defective cable segments and route network traffic around
them. This capability makes locating and repairing bad cable segments easier.
Networks with star wiring topologies can be significantly easier to
troubleshoot and repair than bus-wired networks. With a star network, you can isolate a problem node from the rest of the network
by disconnecting the cable and directly connecting it to the cable
hub. If the hub is considered intelligent, management software
developed for that hub type, as well as the hub itself, can disconnect
the suspect port. Another benefit to this is that one bad cable segment does not affect the entire network, only the machine connected to that bad cable.
10BASE-T enables you to design and build your LAN one segment
at a time, growing as your network needs to grow. This capability
makes 10BASE-T more flexible than other LAN cabling options.
10BASE-T is also relatively inexpensive to use compared to other
cabling options. In some cases in which a data-grade phone system
has already been used in an existing building, the data-grade phone
cable can be used for the LAN.
The rules for a 10BASE-T network are as follows:
á The maximum number of computers on a LAN is 1,024.
á The cabling should be UTP Category 3, 4, or 5. (Shielded
twisted-pair cabling, STP, can be used in place of UTP.)
á The maximum unshielded cable segment length (hub to trans-
ceiver) is 100 meters (328 feet).
175
09.919-x CH04.i
176
8/28/98 12:28 PM
C h a p te r 4
Page 176
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
á The cable minimum distance between computers is 2.5 meters
(8 feet).
á The minimum distance between a hub and a computer, or
between two hubs, is 0.5 meters (1.5 feet).
10BASE-FL
10BASE-FL is a specification for ethernet over fiber-optic cables.
The 10BASE-FL specification calls for a 10Mbps data rate using
baseband.
The advantages of fiber-optic cable (and hence, the advantages of
10BASE-FL) are discussed in Chapter 3. The most important advantages are long cabling runs (10BASE-FL supports a maximum
cabling distance of about 2,000 meters) and the elimination of any
potential electrical complications. Another advantage is that the
number of nodes a segment can handle with 10BASE-FL is far
greater than the maximum supported by 10BASE-T, 10BASE2, and
10BASE5.
100VG-AnyLAN
100VG-AnyLAN is defined in the IEEE 802.12 standard. IEEE
802.12 is a standard for transmitting ethernet and token-ring packets (IEEE 802.3 and 802.5) at 100Mbps. 100VG-AnyLAN is sometimes called 100BASE-VG. The “VG” in the name stands for “voice
grade.” 100VG-AnyLAN cabling uses four twisted-pairs in a scheme
called quartet signaling.
NOTE
The section titled “Demand Priority,” earlier in this chapter, discussed 100VG-AnyLAN’s demand priority access method, which
provides for two priority levels when resolving media access conflicts.
The Upgrade Path Both 100VGAnyLAN and 100BASE-X (see the following section) can be installed as a
Plug and Play upgrade to a 10BASE-T
system.
100VG-AnyLAN uses a cascaded star topology, which calls for a hierarchy of hubs. Computers are attached to child hubs, and the child
hubs are connected to higher-level hubs called parent hubs (see
Figure 4.17).
The maximum length for the two longest cables attached to a
100VG-AnyLAN hub is 250 meters (820 ft). The specified cabling is
Category 3, 4, or 5 twisted-pair or fiber-optic. 100VG-AnyLAN is
compatible with 10BASE-T cabling.
09.919-x CH04.i
8/28/98 12:28 PM
Page 177
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
PARENT HUB
CHILD
HUBS
100BASE-X
FIGURE 4.17
Cascaded star topology.
100BASE-X uses a star bus topology similar to 10BASE-T’s.
100BASE-X provides a data transmission speed of 100Mbps using
baseband.
The 100BASE-X standard provides the following cabling specifications:
á 100BASE-TX. Two twisted pairs of Category 5 UTP or STP.
á 100BASE-FX. Fiber-optic cabling using 2-strand cable.
á 100BASE-T4. Four twisted-pairs of Category 3, 4, or 5 UTP.
100BASE-X is sometimes referred to as Fast Ethernet. Like 100VGAnyLAN, 100BASE-X provides compatibility with existing
10BASE-T systems and thus enables plug-and-play upgrades from
10BASE-T.
In summary, ethernet networks use the following cable types:
á 10BASE2
á 10BASE5
á 10BASE-T
R E V I E W
B R E A K
177
09.919-x CH04.i
178
8/28/98 12:28 PM
C h a p te r 4
Page 178
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
á 10BASE-FL
á 100VG-AnyLAN
á 100BASE-X
• 100BASE-TX
• 100BASE-FX
• 100BASE-T4
Token Ring
Token ring uses a token-passing architecture that adheres to the
IEEE 802.5 standard, as described earlier. The topology is physically
a star, but token ring uses a logical ring to pass the token from
station to station. Each node must be attached to a concentrator
called a multistation access unit (MSAU or MAU).
In the earlier discussion of token passing, it may have occurred to
you that if one computer crashes, the others will be left waiting forever for the token. MSAUs add fault tolerance to the network, so
that a single failure doesn’t stop the whole network. The MSAU can
determine when the network adapter of a PC fails to transmit and
can bypass it.
Token-ring network interface cards can run at 4Mbps or 16Mbps.
Although 4Mbps cards can run at that data rate only, 16Mbps cards
can be configured to run at 4 or 16Mbps. All cards on a given network ring must run at the same rate. If all cards are not configured
this way, either the machine connected to the card cannot have network access, or the entire network can be ground to a halt.
To
ke
n
FIGURE 4.18
Operation of a token ring.
As shown in Figure 4.18, each node acts as a repeater that receives
tokens and data frames from its nearest active upstream neighbor
(NAUN). After the node processes a frame, the frame transmits
downstream to the next attached node. Each token makes at least
one trip around the entire ring and then returns to the originating
node. Workstations that indicate problems send a beacon to identify
an address of the potential failure.
09.919-x CH04.i
8/28/98 12:28 PM
Page 179
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
179
Token Ring Cabling
Traditional token-ring networks use twisted-pair cable. The following are standard IBM cable types for token ring:
>
á Type 1. A braided shield surrounds two twisted pairs of solid
copper wire. Type 1 is used to connect terminals and distribution panels or to connect between different wiring closets that
are located in the same building. Type 1 uses two STPs of
solid-core 22 AWG wire for long, high-data-grade transmissions within the building’s walls. The maximum cabling distance is 101 meters (331 feet).
á Type 2. Type 2 uses a total of six twisted pairs: two are STPs
(for networking) and four are UTPs (for telephone systems).
This cable is used for the same purposes as Type 1, but enables
both voice and data cables to be included in a single cable run.
The maximum cabling distance is 100 meters (328 feet).
á Type 3. Used as an alternative to Type 1 and Type 2 cable due
Type 3 cabling (UTP) is the most popular transmission medium for
token ring. A token-ring network using Type 3 (UTP) cabling can
support up to 72 computers. A token-ring network using STP
cabling can support up to 260 computers.
The minimum distance between computers or between MSAUs is
2.5 meters (8 feet).
A patch cable is a cable that connects MSAUs. Patch cables are typically IBM Type 6 cables that come in standard lengths of 8, 30, 75,
or 150 feet. (A Type 6 cable consists of two shielded 26-AWG
twisted-pairs.) You can also get patch cables in custom lengths.
You can use patch cables to extend the length of Type 3 cables or to
NOTE
to its reduced cost, Type 3 has unshielded twisted-pair copper
with a minimum of two twists per inch. Type 3 has four UTPs
of 22 or 24 AWG solid-core wire for networks or telephone
systems. Type 3 cannot be used for 16Mbps token-ring networks. It is used primarily for long, low-data-grade transmissions within walls. Signals don’t travel as fast as with Type 1
cable because Type 3 doesn’t have the shielding that Type 1
uses. The maximum cabling distance (according to IBM) is 45
meters (about 148 feet). Some vendors specify cabling distances of up to 150 meters (500 feet).
MAU vs. Hub People often confuse
a MAU with a hub. A MAU is strictly
used in a token-ring network. It is
involved with the token generation
and with facilitating the passing of a
token between machines. A hub is
used in ethernet. It is used to connect drop cables from various workstations. These two items are not
interchangeable.
09.919-x CH04.i
180
8/28/98 12:28 PM
C h a p te r 4
Page 180
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
connect computers to MSAUs. Patch cables have an IBM connector
at each end.
Token-ring adapter cables can have an IBM data connector at one
end and a nine-pin connector at the other end, or they can use UTP
cables with RJ-45 connectors on each end. Adapter cables connect
client and server network adapters to other network components
that use IBM data connectors. The type of connectors you need for
a token-ring network depends on the type of cabling you’re using.
Type 3 cabling uses RJ-11 or RJ-45 connectors. (Media filters, if
necessary, can convert the network adapter to RJ-11 or RJ-45 format.) Meanwhile, Type 1 and 2 cabling use IBM Type A connectors.
Token-ring networks come in a few sizes and designs. A small movable token-ring system supports up to 12 MSAUs and uses Type 6
cable to attach clients and servers to IBM Model 8228 MSAUs. Type
6 is flexible but has limited distance capabilities. The characteristics
of Type 6 cable make it suitable for small networks and for patch
cords.
A large nonmovable system supports up to 260 clients and file servers
with up to 33 MSAUs. This network configuration uses IBM Type 1
or Type 2 cable. The large nonmovable system also involves other
wiring needs, such as punch panels or distribution panels, equipment racks for MSAUs, and wiring closets to contain the previously
listed components.
The MSAU is the central cabling component for IBM Token-Ring
networks. The 8228 MSAU was the original wiring hub developed
by IBM for its IBM Token-Ring networks. (IBM names all its hardware with numbers.) Figure 4.19 shows 8228 MSAUs. Each 8228
has ten connectors, eight of which accept cables to clients or servers.
The other connectors are labeled RI (ring in) and RO (ring out).
The RI and RO connectors are used to connect multiple 8228s to
form larger networks. The last RO must be connected to the first
MAU’s RI.
8228s are mechanical devices that consist of relays and connectors.
Their purpose is to switch clients in and out of the network. Each
port is controlled by a relay powered by a voltage sent to the MSAU
from the client. When an 8228 is first set up, each of these relays
must be initialized with the setup tool that is shipped with the unit.
Insert the setup tool into each port and hold it there until a light
indicates that the port is properly initialized.
09.919-x CH04.i
8/28/98 12:28 PM
Page 181
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
150 Feet Maximum
181
FIGURE 4.19
An example of token-ring cabling using MSAUs,
also known as MAUs.
8 Feet Minimum
Patch Cable
RI
RO
RI
RO
Adapter Cable
Token Ring Cabling
When you connect a token-ring network, make sure you do the following:
1. Initialize each port in the 8228 MSAU by using the setup tool
shipped with the MSAU.
2. If you’re using more than one MSAU, connect the RO port of
each MSAU with the RI port of the next MSAU in the loop.
3. Connect the last RO with the first RI to complete the loop so
that the MSAUs form a circle or ring.
Passing Data on Token Rings
As this chapter has already described, a frame called a token perpetually circulates around a token ring (see Figure 4.20). The computer
that holds the token has control of the transmission medium. The
actual process is as follows:
1. A computer in the ring captures the token.
2. If the computer has data to transmit, it holds the token and
transmits a data frame. A token-ring data frame contains the
fields listed in Table 4.1.
3. Each computer in the ring checks to see whether it is the
intended recipient of the frame.
09.919-x CH04.i
182
8/28/98 12:28 PM
C h a p te r 4
Page 182
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
4. When the frame reaches the destination address, the destination PC copies the frame to a receive buffer, updates the frame
status field of the data frame (see step 2), and puts the frame
back on the ring.
In 16Mbps token-ring networks, the sending device can utilize
an optional enhancement, known as early token release. This is
where the sending device issues a token immediately after
sending a frame, not waiting for its own header to return. This
speeds up the data transfers on the network.
5. When the computer that originally sent the frame receives it
from the ring, it acknowledges a successful transmission, takes
the frame off the ring, and places the token back on the ring.
TABLE 4.1
T O K E N - R I N G D ATA F R A M E F I E L D S
Start
deliminator
FIGURE 4.20
A token ring frame.
Access
control
Frame
control
Dest.
Address
Field
Description
Start delimiter
Marks the start of the frame
Access control
Specifies priority of the frame; also specifies whether the
frame is a token or a data frame
Frame control
Media Access Control information
Destination address
Address of receiving computer
Source address
Address of sending computer
Data
Data being transmitted
Frame check sequence
Error-checking information (CRC)
End delimiter
Marks the end of the frame
Frame status
Tells whether the destination address was located and
whether the frame was recognized
Source
Address
Data
Frame
check
sequence
End
deliminator
Frame
status
09.919-x CH04.i
8/28/98 12:28 PM
Page 183
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
The Beaconing Process
Generally, the first station that is powered up on a token-ring network automatically becomes what is called the active monitor station.
The responsibility of the active monitor station is to announce itself
to the next active downstream station as the active monitor station
and request that station to announce itself to its next active downstream station. The active monitor station sends this beacon
announcement every seven seconds.
After each station announces itself to its next active downstream
neighbor, the announcing station becomes the nearest active
upstream neighbor (NAUN) to the downstream station. Each station on a token-ring network has an upstream neighbor as well as a
downstream neighbor.
After each station becomes aware of its NAUN, the beaconing
process continues every seven seconds. If, for some reason, a station
doesn’t receive one of its expected seven-second beaconed announcements from its upstream neighbor, it attempts to notify the network
of the lack of contact from the upstream neighbor. It sends a message out onto the network ring, which includes the following:
á The sending station’s network address
á The receiving NAUN’s network address
á The beacon type
From this information, the ring can determine which station might
be having a problem and then attempt to fix the problem without
disrupting the entire network. This process is known as autoreconfiguration. If autoreconfiguration proves unsuccessful, there may be a
Ring Purge issued by the active monitor, forcing all computers to
stop what they are doing and resynchronize with the ring. If both
these mechanisms fail, manual correction becomes necessary. Figure
4.21 shows a token-ring network utilizing the beaconing process.
ARCNet
ARCNet is an older architecture that is not found too often in the
business world, but does have a presence in many older networks
and school systems who often receive hand-me-downs from the
business sector.
183
09.919-x CH04.i
184
8/28/98 12:29 PM
C h a p te r 4
FIGURE 4.21
Token-ring beaconing.
Page 184
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
1
2
C:>
station 1
C:>
station 2
Active Monitor
station 1
station 2
Active Monitor
C:>
station 3
station 4
Station 2 powers up, sends out-of-frame
to next powered up station.
3
station 3
station 4
Station 4 powers up, receives station 2's
out-of-frame, introduces itself to next
powered up station, station 3.
4
F:>
station 1
F:>
station 3
station 2
Active Monitor
F:>
station 4
Station 3 powers up, receives station 4's
introduction and request to introduce itself
to next powered up station, station 1.
F:>
station 1
F:>
station 2
Active Monitor
F:>
station 3
F:>
station 4
Station 1 powers up, receives
station 3's introduction.
ARCNet utilizes a token-passing protocol that can have a star or bus
physical topology. These segments can be connected with either
active or passive hubs. ARCNet, when connected in a star topology,
can use either twisted pair or coaxial cable (RG-62). If coaxial cable
is used to create a star topology, the ends of the cable can be
attached directly to a BNC connector, without a terminator. When
in a bus topology, ARCNet uses a 93-ohm terminator, which is
attached to each end of the bus in a similar fashion to an ethernet
bus.
Each ARCNet card has a set of DIP switches built onto it. You can
change the setting of these DIP switches to give each card a separate
hardware address. (This is covered in more detail in Chapter 5.)
Based upon these addresses, tokens are passed to the card with the
next highest address on the network. Due to this “access to the network passing,” ARCNet shares some characteristics with a token
passing network.
Some important facts about ARCNet are as follows:
09.919-x CH04.i
8/28/98 12:29 PM
Page 185
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
á ARCNet uses a 93-ohm terminator. (Ethernet uses a 50-ohm
terminator.)
á ARCNet uses a token-like passing architecture, but does not
require a MAU.
á The maximum length between a node and an active hub is
610 meters. (Hubs are discussed in more detail in Chapter 5,
“Connecting Devices.”)
á The maximum length between a node and a passive hub is
30.5 meters.
á The maximum network segment cable distance ARCNet sup-
ports is 6100 meters.
á ARCNet can have a total of only 255 stations per network
segment.
FDDI
FDDI is very similar to token ring in that it relies on a node to have
the token before it can use the network. It differs from token ring in
that it utilizes fiber-optic cable as its transmission media, allowing
for transmissions of up to 100Km. This standard permits up to 100
devices on the network with a maximum distance between stations
of up to 2 kilometers (see Figure 4.22).
FDDI has two different configurations: Class A and B. Class A uses
two counteracting rings. Devices are attached to both rings. If one
of these rings develops a fault, the other ring can still be used to
transmit data. Class B uses a single ring to transmit data.
Ring 1
FIGURE 4.22
Ring 2
An FDDI Network.
185
09.919-x CH04.i
186
8/28/98 12:29 PM
C h a p te r 4
Page 186
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
C A S E S T U DY : S M A L L -C O M PA N Y T O P O L O G Y
ESSENCE OF THE CASE
These are the essential facts:
• Fast access to the database server is
needed by all parties.
• Data is transferred in very small
amounts, but access on the network is
constant by almost all 100 employees.
Bookings
Check in
Staff
• The company wants to minimize the
effects of transmission media failure on
the network.
Admin Staff
• Walls between groups cannot be drilled
through, so all cable must go around the
wall.
Server
FIGURE 4.23
SCENARIO
You are responsible for deciding upon a topology
for a small airline company’s network. This company has 80 employees. The physical layout of
the building that this airline company is located
in is seen in Figure 4.23.
The information that flows through this company
consists mostly of airline reservations. The company is using a Microsoft NT server with a SQL
server database running on it. The database is
accessed by each of three groups in this Airline.
These are: Group 1—Reservation and Booking,
Group 2—Check-in, and Group 3—Administrative.
Each group is separated from each of the others
by an impenetrable wall, with no access over or
through the wall. The only way to physically move
anything from one group to another is to go
around the wall. The greatest distance between
the server and a PC is 80 meters.
Physical layout
All three groups have important functions, and no
one group can really afford to have delays in
receiving information from the server. The
Reservation and Bookings staff cannot afford to
have the clients wait to book a flight, which may
cause lost sales. They also must be able to
process persons waiting at the terminal or else
flights leave late. Thus they need quick access to
the database server. The check-in staff has to be
able to quickly access the database on the server to check in customers, so that flights will not
be delayed. The Administrative staff has to be
able to access the database server quickly to
deal with complaints and lost luggage. Because
the airline does not run flights constantly, a lot of
people tend to be on the network at one time, or
very few. Also, the amount of data transferred on
the network by each user is relatively small (how
09.919-x CH04.i
8/28/98 12:29 PM
Page 187
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
187
C A S E S T U DY : S M A L L - C O M PA N Y T O P O L O G Y
much information is really on an airline ticket?),
but all the users are continually accessing the
network when the flights are running. Only a
handful would be continually accessing the network during those times that the flights are not
running.
Because this is a competitive market, this airline
company cannot afford to have the network go
down.
A N A LY S I S
This case study is a good example of how to deal
with a real-world issue as well as how to address
the exam topic of “Select the appropriate topology for various token-ring and ethernet networks.”
One of the first questions needing to be asked is
whether this network should use contention
access or token passing. Fast access to the
server is needed by all parties in question. Both
contention-based systems and token-ring systems can give fast access to the network, but
as the network load increases, one would find
that a token-ring system would start to perform
better. If the load on the system is very high, the
token-ring solution is the best, because it guarantees all parties equal access to the network at
once. The drawback to token ring is that it has a
higher price tag than that of ethernet.
If the decision is made to use token ring, the
topology of choice is going to be a physical star.
A physical star is nice, in that if a patch cable is
broken, this does not affect other nodes on the
network. If the decision is made to go with ethernet, there are two options for the physical topology: a bus or a star. The benefits of a physical
star are the same for ethernet as for token-ring.
A bus physical topology has the disadvantage
that if any segment is broken, the entire network
segment fails, affecting all parties. Furthermore,
this point of failure is difficult to trace. Thus, due
to the requirements of the company, in that they
want to minimize the effects on the network due
to failure of the transmission media, a bus topology should be ruled out.
To this point, your options have come down to
the physical star ethernet or the token ring. The
final issue to resolve is the type of cable to use
in each situation.
For the ethernet solution, the 10BASE2 and
10BASE5 are ruled out, because they are used in
a bus topology. 10BASE-T meets the requirements, because it supports cable runs of up to
100 meters, and you are not be going farther
than 80 meters. Both 100BASE-X and 100BASEAnyLAN can also be used, but these alternatives
cost more than the 10BASE-T, and the data
amounts transmitted are very small, hence there
is probably not a strong reason to justify the
extra cost.
For the token-ring network, Type 1 and 2 cables
would work. The IBM standards for Type 3 do not
allow the cable run to go past 45 meters; it is
therefore too short. In addition, Type 3 does not
support 16Mbps. For this reason, even if an
approved vendor version that supported lengths
up to 150 meters were used, Type 3 would probably be too slow.
To narrow down the cable type even further, in
the Token-Ring network, Type 2 cable is more
expensive than Type 1, hence, the preferred
option is Type 1.
continues
09.919-x CH04.i
188
8/28/98 12:29 PM
C h a p te r 4
Page 188
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
C A S E S T U DY : S M A L L -C O M PA N Y T O P O L O G Y
continued
In summary, the following two solutions have
been proposed:
• Ethernet running 10BASE-T
• Token ring running Type 1 cable
Between these options there is no correct
answer. Token ring costs more than the ethernet
solution, but at the same time gives faster
access to the users during peak usage. The company would need to evaluate whether the higher
cost is worth the better performance of the network during peak time usage. A smart next step
would be to see what other airlines of the same
size are experiencing when they use either one of
the topologies.
CHAPTER SUMMARY
KEY TERMS
• Contention
• Polling
• Token passing
• Physical topology
• Logical topology
• Bus topology
• Ring topology
• Star topology
• Mesh topology
• Ethernet
This chapter examined some common network topologies. You
learned about the basic access methods, such as contention, polling,
and token passing. This chapter then described some fundamental
topology archetypes (bus, ring, and star) and discussed the differences between physical and logical topologies. Lastly, the chapter
described the common varieties of ethernet and token-ring networks, as well as information on ARCNet and FDDI networks.
Just as you did in Chapter 2, you should review this chapter in its
entirety. Questions concerning different network topologies are
often stated in terms of comparing the features between the two
topologies.
In short, ethernet topologies
• Token ring
á Are contention based
• ARCNet
á Often employ a 10BASE or 100BASE cable types
• FDDI
á Can be employed in a physical star or bus topology
• 10BASE2
• 10BASE5
• 10BASE-T
• 10BASE-FL
• 100VG-AnyLAN
• 100BASE-X
Token-ring networks
á Use a token passing access method
á Use an IBM type of cabling standard
á Mostly use a physical star topology
09.919-x CH04.i
8/28/98 12:29 PM
Page 189
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
189
A P P LY Y O U R L E A R N I N G
Exercises
4.1
Matching Topologies to Applications
Objective: Practice associating network topologies with
appropriate uses.
Time estimated: 10 minutes.
Match the topology to the application. For this exercise, you should be familiar with the material in this
chapter and also in Chapter 3.
1. 10BASE2
2. 10BASE5
3. 10BASE-T
4. 10BASE-FL
5. 100BASE-X
6. Token Ring
A. You are looking for an inexpensive network with the
maximum flexibility for future
expansion. You want to utilize
existing data-grade phone
lines for some segments.
B. Your network encompasses
three buildings. The longest
segment length is 450 meters.
You want to minimize cost.
The difference in electrical
ground potential between the
buildings is not a problem.
C. Your company encompasses
three buildings. The longest
segment length is 1,800
meters. In previous networking attempts, you have experienced problems with the
ground potential differences
between the buildings.
D. You are designing a network
for an airline ticket office.
Employees query the database
constantly, so the network utilization rate is extremely high.
The network must be very reliable and capable of
self-corrective action to isolate a malfunctioning
PC.
E. You work in a small office with 12 PCs. You are
looking for an inexpensive networking solution.
The computers are spaced evenly throughout the
office (approximately 3–5 meters between workstations). You want to minimize the total amount
of cabling.
F. Your company colorizes Hollywood movies.
Huge, digitized movie files, such as Bringing Up
Baby or The Jazz Singer, must pass quickly
through the network so they arrive with extreme
dispatch at colorizing workstations. Very high
transmission speeds are required. Your company
is reaping huge profits, so the cost of cabling is
no concern.
The correct responses are as follows:
1. E
2. B
3. A
4. C
5. F
6. D
Review Questions
1. Explain the difference between a logical and a
physical topology.
2. Explain the difference between a physical bus
topology and a physical star topology.
3. Explain the difference between contention-based,
polling, and token-passing access methods.
09.919-x CH04.i
190
8/28/98 12:29 PM
C h a p te r 4
Page 190
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
A P P LY Y O U R L E A R N I N G
Exam Questions
1. CSMA/CD uses which two of the following techniques to control collisions?
A. Nodes broadcast a warning before they transmit.
C. Token-ring
D. IEEE 802.3
5. If you see a group of networked computers connected to a central hub, you know that the network has what type of physical topology?
B. Nodes listen for a clear line before they transmit.
A. Ring
C. Nodes request and are given control of the
medium before transmitting.
C. Bus
D. Nodes listen while they transmit and stop
transmitting if another signal interferes with
the transmission.
2. What is the maximum size of a 10BASE5 network?
B. Star
D. Can’t tell
6. If you see a group of networked computers connected to a central concentrator, you know that
the network has what type of logical topology?
A. Ring
A. 100 meters
B. Star
B. 300 meters
C. Bus
C. 1,500 meters
D. Can’t tell
D. 2,500 meters
3. By what type of network is CSMA/CA commonly used?
A. Microsoft networks
B. LocalTalk networks
7. Which topology uses fiber-optic cable?
A. 10BASE2
B. 10BASE5
C. 10BASE-T
D. None of the above
C. Fast Ethernet networks
D. 10BASE5 networks
8. Which topology uses Thicknet cable?
A. 10BASE2
4. Which three of the following network architectures use the token-passing access method?
A. IEEE 802.4
B. FDDI
B. 10BASE5
C. 10BASE-T
D. None of the above
09.919-x CH04.i
8/28/98 12:29 PM
Page 191
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
191
A P P LY Y O U R L E A R N I N G
9. Which topology uses UTP cable?
A. 10BASE2
B. 10BASE5
C. 10BASE-T
D. None of the above
10. Which topology uses Thinnet cable?
A. 10BASE2
B. 10BASE5
C. 10BASE-T
D. None of the above
C. 100VG-AnyLAN
D. 100BASE-X
14. A token-ring network using STP cabling can support how many computers?
A. 60
B. 260
C. 500
D. 1,024
15. What field of a token-ring frame is updated by
the destination PC?
A. Destination address
11. 10BASE5 networks consisting of a single cable
segment cannot exceed what maximum length?
A. 185 meters
B. 300 meters
C. 500 meters
D. 1,000 meters
12. Which two of the following are characteristics of
a 10BASE-T network but not a 10BASE2 network?
A. T-connector
B. Central hub
C. UTP
D. BNC
13 What is sometimes called “Fast ethernet”?
B. Frame check sequence
C. End delimiter
D. Frame status
16. Which two of the following statements are true?
A. Coax ethernet is a physical bus and a logical
bus.
B. 10BASE-T ethernet is a physical bus and a
logical bus.
C. Coax ethernet is a physical star and a logical
bus.
D. 10BASE-T ethernet is a physical star and a
logical bus.
17. What is the single biggest advantage of using
10BASE-T when network segments don’t have to
exceed 185 meters?
A. 10BASE-T
A. It is relatively simple to connect.
B. 10BASE5
B. Drop cables can be used, making it easier to
troubleshoot.
09.919-x CH04.i
192
8/28/98 12:29 PM
C h a p te r 4
Page 192
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
A P P LY Y O U R L E A R N I N G
C. Each node connects directly to the coaxial
cable.
A. This solution meets the primary objective and
both secondary objectives.
D. It is the least expensive of the cabling options.
B. This solution meets the primary objective and
one secondary objective.
18. Which three of the ethernet topologies require
that each end of the bus be terminated?
A. 10BASE2
C. This solution meets the primary objective.
D. This solution does not meet the primary
objective.
B. 10BASE5
C. 10BASE-T
D. 10BASE-FL
19. Which of the following is not an advantage of
using 10BASE-T for cabling a network?
A. It is easier and more reliable to manage.
B. Centralized hubs make it easier to detect bad
cable segments.
C. Beaconing helps to isolate cable breaks.
D. It is relatively inexpensive to use.
20. You are required to select a topology for the corporate network’s backbone. Your decision, which
is to look at only inter-server (not workstation)
connectivity, needs to reflect speed and fault tolerance.
Primary Objective: The topology needs to have
total fault tolerance, in case of a cable break.
21. You wish to install a network that can handle
large traffic volumes.
Primary Objective: The topology needs to be
deterministic. That is, you need to be able to calculate the effect that the addition of new computers will have on the network.
Secondary Objective: Each station is not generating large amounts of data, but will be continuously generating small amounts of data. The
topology needs to be capable of evenly distributing access time to the network.
Secondary Objective: The topology does not need
to accommodate more than 244 computers.
Suggested Solution: Implement the network
using a token-ring network.
A. This solution will obtain the required result
and both optional results.
B. This solution will obtain the required result
and one of the optional results.
Secondary Objective: Large data throughput is
required.
C. This solution will obtain the required result.
Secondary Objective: The topology does not need
to span a distance of more than 25 meters.
D. This solution does not satisfy the required
result.
Suggested Solution: Implement the network
using fiber-optic cabling.
09.919-x CH04.i
8/28/98 12:29 PM
Page 193
Chapter 4
NETWORK TOPOLOGIES AND ARCHITECTU R E S
193
A P P LY Y O U R L E A R N I N G
Answers to Review Questions
1. A logical topology is the logical path that a signal
follows on the transmission media. A physical
topology is the physical layout of the transmission media. See “Network Topologies.”
2. A physical bus is a cable segment that is a straight
line, although in actual practice this cable segment is often snaked around its surroundings.
A physical star has a center hub, with cable segments running out from the hub to the devices
connected to the hub. See “Bus Topologies” and
“Star Topologies.”
3. Contention-based access methods employ either a
CSMA\CD or a CSMA\CA method. CSMA\CD
accesses the network without regard to other
devices on the network. CSMA\CA polls the network first to see whether the media is currently
busy.
The polling access method polls different devices
to see whether they are waiting to transmit or
receive information. A computer accessing a
printer often does this.
Token-passing requires a device to be in possession of a token before transmitting data onto the
network. See “Access Methods.”
2. D. A 10BASE5 network is capable of spanning
up to 500 meters on a single segment. Because
you can add repeaters, and based on the 5-4-3
rule (5 cable segments connected by 4 repeaters
and only three of these segments can be populated), a 10BASE5 network can span up to 2500
meters. See the section titled “10BASE5.”
3. B. Apple’s LocalTalk networks utilize this contention mechanism. See the section titled
“Contention” under the topic of “Access
Methods.”
4. A, B, C. D is an ethernet standard defined by the
IEEE. See the section titled “Token Passing”
under the topic of “Access Methods.”
5. B. A is a logical topology and B does not use a
hub for connecting. See the section titled “Star
Topologies” under the topic of “Physical and
Logical Topologies.”
6. D. The logical topology could either be a ring or
bus. See the topic titled “Physical and Logical
Topologies.”
7. D. 10BASE-FL is fiber-optic. See the topic titled
“Ethernet.”
8. B. A is coaxial Thinnet cable and C is twistedpair cable. See the topic titled “Ethernet.”
9. C. C is twisted pair, which is what UTP is. See
the topic titled “Ethernet.”
Answers to Exam Questions
1. B, D. A is CSMA/CA, and C is indicative of
token ring. The purpose of CSMA/CD is to listen before transmitting and then stop if a collision occurs. See the section titled “Contention”
under the topic of “Access Methods.”
10. A. 10BASE2 is often referred to as Thinnet cable.
See the topic titled “Ethernet.”
11. C. Thicknet cables can go only 500 meters before
needing a repeater. See the topic titled
“Ethernet.”
09.919-x CH04.i
194
8/28/98 12:29 PM
C h a p te r 4
Page 194
NE TWO R K TO P O L OGIES AND ARCHITECTU RES
A P P LY Y O U R L E A R N I N G
12. B, C. A and D are used on coaxial cable. Twistedpair cable goes into a hub, and another term for
twisted-pair cable is Unshielded Twisted Pair
(UTP). See the topic titled “Ethernet.”
13. D. This is the other term for 100BASE-X. See
the topic titled “Ethernet.”
14. B. The maximum number of devices for a token
ring using STP is 260. See the topic titled
“Token Ring.”
15. D. A, B, and C are either generated by the source
PC or read by the destination PC. See the topic
titled “Token Ring.”
16. A, D. B is incorrect because it says “physical bus,”
C is incorrect because it says “physical star.” See
the topic titled “Ethernet” and the topic titled
“Token Ring.”
17. D. A could be correct, but this would be dependent upon the network layout; therefore this is
not always an advantage. B is incorrect, because
drop cables are not used. C is incorrect, because
10BASE-T does not use coaxial cable. See the
topic titled “Ethernet.”
18. A, B, D. Terminators are not used in 10BASE-T.
See the topic titled “Ethernet.”
19. C. Beaconing is done on a token-ring network.
See the topic titled “Token Ring.”
20. D. The only topology that supports total fault
tolerance is a mesh topology, where every computer is connected to every other computer, using
an independent connection. A fiber-optic network could be the transmission media of choice
and does both secondary objectives, but does not
satisfy the primary objective. See the section
titled “Network Topologies.”
21. A. The primary objective and both secondary
objectives are indicative of a token ring network.
See the section titled “Token Ring.”
Suggested Readings and Resources
1. Tannenbaum, Andrew. Computer Networks.
Prentice-Hall, 1996.
2. Derfler, Frank, Jr. and Les Freed. How
Networks Work. Ziff-Davis Press, 1996.
10.919-x CH05.i
8/28/98 12:55 PM
Page 195
OBJECTIVES
Chapter 5 targets the following objectives in the
Implementation section of the Networking Essentials
exam (it has been included here in the Planning part to
be consistent with the organization of the chapters
reflecting the layers of the OSI model):
Given the manufacturer’s documentation for the
network adapter, install, configure, and resolve
hardware conflicts for multiple adapters in a
token-ring or ethernet network.
. As a network professional, it is vital that you have
the ability to not only install and configure network
adapter cards, but also to resolve hardware conflicts
between network adapter cards and other adapters
and peripherals. If these conflicts exist and are not
resolved, your network adapter card will not function, thereby preventing your device from connecting to the network.
C H A P T E R
5
Network Adapter
Cards
10.919-x CH05.i
8/28/98 12:55 PM
Page 196
OUTLINE
Defining the Workings of a Network
Adapter Card
S T U DY S T R AT E G I E S
197
. The topic of network adapters requires you to
know how to accomplish the following tasks:
Preparing and Sending Data
198
• Installing a network adapter card
How a Network Card Works
Signals
Clocking
Measurement of the Signal
199
199
201
202
• Configuring a network adapter card
Installing Network Adapter Cards
204
Configuring Network Adapter Cards
206
IRQ
207
Base I/O Port Address
207
Base Memory Address
208
DMA Channel
208
Boot PROM
208
MAC Address
208
Ring Speed
209
Connector Type
209
Resolving Hardware Conflicts
209
Chapter Summary
213
• Resolving hardware conflicts
. While there is factual knowledge to be learned
concerning this objective, make sure you work
your way through the Step-by-steps and
Exercises in this chapter. That kind of hands on
experience is important to answering exam
questions on this topic.
10.919-x CH05.i
8/28/98 12:55 PM
Page 197
C hapter 5
INTRODUCTION
When devices are attached to a network, some mechanism must
exist for transferring the information from one device to a transmission medium so that the other device or devices on the network can
receive the information. Likewise, the receiving device must also
have some mechanism to receive this information from the transmission medium, so that it can process the information. This chapter
and Chapter 6, “Connectivity Devices and Transfer Mechanisms,”
describe to you the function and workings of some of the common
devices used to attach components to the network’s transmission
medium. This chapter examines the role of the network adapter card,
also known as a network interface card (NIC). Because a network
adapter card is the most common mechanism for attaching PCs to a
network, it is deserving of an entire chapter. The following chapter,
Chapter 6, explains some of the other more common devices used
to transmit data in a network.
A network adapter card is a hardware device that installs in a PC
and provides an interface from a PC to the transmission medium.
Most PC networks, including ethernet, token-ring, and ARCNet,
use network adapter cards. The network adapter card is thus an
essential part of networking, and an understanding of network
adapter cards is crucial for any networking professional.
This chapter begins by explaining the mechanisms used to transfer
data from the PC to the transmission medium. From there, configurable options on a network card are examined. That section
explains the different properties that are configurable on a network
adapter card. The next section of the chapter goes on to explain the
installation of a network adapter card, and the chapter concludes
with ways to resolve hardware conflicts.
DEFINING THE WORKINGS OF
NETWORK ADAPTER CARD
A
A network adapter card links a PC with the network cabling system
(see Figure 5.1). The network adapter card fits into one of the PC’s
expansion slots. The card has one or more user-accessible ports to
which the network cabling medium is connected.
NETWORK ADAPTER CAR D S
197
10.919-x CH05.i
198
8/28/98 12:55 PM
C h a p te r 5
Page 198
NE TWO R K ADAP TER CARDS
FIGURE 5.1
I/0
Memory
An example of a network adapter card.
Disable
DC000
D8000
CC000
C8000
300
310
330
350
250
280
2A0
2E0
Socket
for Remote
Boot PROM
Memory Address Jumpers
I/O Address Jumpers
BNC Connector
"DIX" Connector
Network adapter cards play an important role on the network. They
are responsible for translating data from a device on the network—
mostly computers—and converting this data into some form of signal that can be transmitted across the transmission medium. To
enable you to understand the full functionality of the network
adapter card, the specific functions of what the network card does
must each be addressed.
Preparing and Sending Data
All network cards perform the function of preparing and sending
data from a computer to the transmission medium. This data, when
inside the computer, travels along the bus of a computer in parallel
form. This data can move at 8, 16, or 32 bits at a time. The network
card must convert these signals coming to it in parallel form, into a
serial signal that can travel across the transmission medium.
Likewise, when data is received, this serial form of data that is in the
signal must be converted into a parallel form matching the bus type
(8, 16, or 32 bit) being used by the receiving device.
DATA BUS E S
The data bus is a pathway inside your computer that carries data
between the hardware components. Four data bus architectures are
used in Intel-based PCs: the 8-bit bus or Industry Standard
Architecture (ISA), the 16-bit bus Extended Industry Standard
10.919-x CH05.i
8/28/98 12:55 PM
Page 199
C hapter 5
Architecture (EISA), the 32-bit bus known as the Micro Channel
(MCA), and the 32-bit bus also known as the Peripheral Component
Interconnect (PCI). In recent models, PCI and EISA are the most
common data bus architectures. ISA is a (more limited) predecessor of EISA. Micro Channel is a data bus developed by IBM for the
PS/2 series that never caught on because it needed to be licensed
from IBM by other equipment manufacturers.
The mechanism of this data conversion is handled in two ways.
First, when data is coming from the computer, to be prepared to be
sent out on the network, the network adapter card’s driver, or software interface, is responsible for converting this data into a format
that can be understood by the network adapter card. As explained in
Chapter 2, “Networking Standards,” this standard was either NDIS
or ODI, depending on whether you were going to interface with a
Microsoft operating system (NDIS) or a Novell operating system
(ODI).
The second part of the data conversion is performed by the physical
network card itself. It is here that the actual data that has been
passed along from the computer is converted into a serial format
using either a digital, analog, or light signal. The network card not
only converts the data into this signal, but it also is responsible for
accessing the transmission medium and forming a channel to conduct the signals onto the network. In essence, a network card is like
the doorway to the network for the PC or other device.
How a Network Card Works
To enable you to fully appreciate how a network card functions, two
important concepts must be explained. These are signals and clocking.
Signals
Two basic types of signals are used with transmission media: analog
and digital.
Analog Signals
Analog signals (seen in Figure 5.2) constantly vary in one or more
values, and these changes in values can be used to represent data.
Analog waveforms frequently take the form of sine waves.
NETWORK ADAPTER CAR D S
199
10.919-x CH05.i
200
8/28/98 12:55 PM
C h a p te r 5
Page 200
NE TWO R K ADAP TER CARDS
FIGURE 5.2
An example of an analog signal.
ANALOG
SIGNAL
The two characteristics that define an analog waveform are as follows:
á Frequency. Indicates the rate at which the waveform changes.
Frequency is associated with the wavelength of the waveform,
which is a measure of the distance between two similar peaks
on adjacent waves. Frequency generally is measured in Hertz
(Hz), which indicates the frequency in cycles per second.
Frequency is illustrated in Figure 5.3.
á Amplitude. Measures the strength of the waveform. Amplitude
is illustrated in Figure 5.4.
Each of these characteristics—frequency and amplitude—can be
used to encode data.
0
Time
FIGURE 5.3
These two analog waveforms differ in frequency.
Higher Frequency
Lower Frequency
10.919-x CH05.i
8/28/98 12:56 PM
Page 201
C hapter 5
NETWORK ADAPTER CAR D S
FIGURE 5.4
These two waveforms differ in amplitude.
0
Time
Higher Amplitude
Lower Amplitude
Digital Signals
Digital signals are different than analog signals in that digital signals
have two discrete states. These states are either “off ” or “on.” An
example of how a digital signal is represented is seen in Figure 5.5.
Clocking
Clocking is the mechanism used to count and pace the number of
signals being sent and received. Signals are expected to be sent in a
continuous flow, representing the start and ending of the data.
Clocking is the mechanism used by the network adapter card to
determine how much data has been sent. For example, if a network
card is designed to transmit data at 20,000 Megahertz a second,
other cards receiving this data will also read the data at 20,000MHz
a second. Clocking is a mechanism used by all network adapter cards
to measure how much data has been sent or received.
A good example of clocking is when a person taps his feet to keep
the time to music. The person doing the tapping expects a set number of music beats per measure; computer network cards also expect
so many signals per second.
A clocking mechanism used by some network cards is oversampling.
With oversampling, the receiving network adapter card samples, or
DIGITAL
SIGNAL
FIGURE 5.5
An example of a digital signal.
201
10.919-x CH05.i
202
8/28/98 12:56 PM
C h a p te r 5
Page 202
NE TWO R K ADAP TER CARDS
reads the signals, at a higher frequency than that at which the data is
sent. This capability is programmed into the card by the manufacturers because the clock used on the sending adapter card can drift
apart from that of the receiving adapter card. Oversampling enables
the clocking mechanism to determine when this drifting apart is
happening so that it can correct the clocking rates.
Measurement of the Signal
To this point you are aware that a network card transmits data and
that this data is transmitted between devices across some transmission medium. The network adapter card’s role is to convert data
from one PC to signals, or convert signals back to understandable
data for the PC. These signals are either analog or digital. You also
know that clocking is used to count the signals. The last step to
understand is the mechanism used by the network adapter card to
read the signals. It should be no big surprise at this point that the
mechanisms can be grouped into two common methods: digital and
analog.
Measurement of Digital Signals
Digital signals use one of two common measurement mechanisms:
current state or state transition. The manufacturer builds these measurement capabilities into the network adapter card.
Current State
Current state is a mechanism that uses the clock count to analyze the
current state of the signal during that count. Thus the signal is either
“on” or “off ” during the clock count. Figure 5.6 shows the idea of
current state measurement.
As seen in Figure 5.6, during each count, the state of the digital signal is either “on” or “off.” Changes in the voltages happen during
changes in the count. Common digital signal schemes for this sampling mechanism are also known as Polar, Unipolar, and Biphase.
State Transition
State transition is a more common form of data measurement of digital signals. This form of measurement is used on ethernet networks
utilizing copper cables. This form of measurement tends to be less
10.919-x CH05.i
8/28/98 12:56 PM
Page 203
C hapter 5
NETWORK ADAPTER CAR D S
FIGURE 5.6
Data
Data
1
1
1
1
1
0
1
0
0
Time in thousandths of a second
prone to signal disruptions and also does not rely as much on the
strength of a signal.
State transition relies on the change of the state of a network signal
to represent a new transmission of data. Recall that in current state
the length of time a signal is on or off indicates whether the signal
represents a 1 or a 0. State transition represents a 1, for example,
every time the state of the signal changes on a count, but a 0 is represented every time the state of the signal does not change during a
count.
Common state transmission measurement standards are Manchester,
Differential Manchester, and Biphase Space.
Measurement of Analog Signals
Analog, much like digital, signals also follow a similar mechanism of
measurement of signals. The main difference between digital and
analog signals is that digital signals have two discrete states—“on”
and “off ”—and analog signals can change frequencies.
Current State
Two mechanisms using current state measurement technologies are
the Frequency Shift Keying (FSK) and Amplitude Shift Keying
(ASK). FSK uses a change in frequency to indicate a change in data,
whereas ASK uses a change in amplitude to indicate a change in
data. An example of Frequency Shift Keying is shown in Figure 5.7.
Current state measurement.
1
203
10.919-x CH05.i
204
8/28/98 12:56 PM
C h a p te r 5
Page 204
NE TWO R K ADAP TER CARDS
FIGURE 5.7
Frequency shift keying and amplitude shift
keying.
1
0
0
1
1
1
0
FSK
0
0
ASK
State Transition
State transition of a frequency is the measurement of a frequency’s
phase during a clock count. A phase is a difference in transition of a
frequency. The transition of a frequency is the change between two
frequencies. Figure 5.8 illustrates this.
An example of phase measurement is that a 1 may be represented by
a 90 degree phase shift, and a 0 by no phase shift.
FIGURE 5.8
These two analog forms differ in phase.
0
90°
Wave A
Wave B
Time
INSTALLING NETWORK ADAPTER
CARDS
Given the manufacturer’s documentation for the network adapter,
install, configure, and resolve hardware conflicts for multiple network adapters in a token-ring or ethernet network.
The details of how to install a network adapter card might depend
on the card, the operating system, or the hardware platform, but the
steps are basically the same. To install a network adapter card, you
must follow these steps:
1. Physically plug the card into the expansion slot, configuring
jumpers and DIP switches as required (see the next section).
2. Install the network adapter card driver.
10.919-x CH05.i
8/28/98 12:56 PM
Page 205
C hapter 5
NETWORK ADAPTER CAR D S
3. Configure the operating system so that the network adapter
card doesn’t conflict with other devices (see the next section).
4. Bind the network adapter to the required protocols (see
Chapter 7, “Transport Protocols,” for more information).
5. Attach the network cable to the card.
Depending on whether the network adapter card’s hardware is plugand-play and if the operating system being used is also plug-andplay, some of these steps might happen automatically when you plug
a card into the slot and start your system. Windows NT is not really
plug-and-play-capable, so when you install a network adapter card
after the operating system is in place, you might have to spend some
time with steps 2–4. Be warned, though: Even the presence of plugand-play devices on plug-and-play systems does not guarantee the
automation of installing hardware.
To install a network adapter card driver in Windows NT, follow
these steps:
STEP BY STEP
5.1 Installing a Network Adapter Card Driver in
Windows NT
1. Click the Start button and choose Settings/Control Panel.
Double-click the Control Panel Network application. In
the Control Panel Network application, choose the
Adapters tab (see Figure 5.9).
FIGURE 5.9
The Control Panel Network Adapters tab.
2. In the Adapters tab (refer to Figure 5.9), click the Add
button to invoke the Select Network Adapter dialog box
(see Figure 5.10). Choose the adapter model from the list
or click the Have Disk button to install a driver that isn’t
listed. Windows NT asks for the location of the Windows
NT installation CD-ROM.
3. Windows NT attempts to detect the adapter and then
might prompt you for additional information (see the section titled “Configuring Network Adapter Cards” later in
this chapter).
FIGURE 5.10
continues
The Select Network Adapter dialog box.
205
10.919-x CH05.i
206
8/28/98 12:56 PM
C h a p te r 5
Page 206
NE TWO R K ADAP TER CARDS
continued
4. When the installation is complete, shut down Windows
NT and restart.
NOTE
NOTE
NOTE
5. Use the Network application’s Bindings tab to check and
set protocol bindings for the new adapter (see Chapter 7).
Make sure the adapter is compatible
with your version of Windows NT. To
do so, check the Windows NT
Hardware Compatibility list or consult
the manufacturer.
The Relationship Between Data Bus
and Processor Type The data bus
architecture is generally independent
of the processor type. Two Pentium
machines from different vendors
might have different data bus architectures, although PCI has essentially
become the standard.
Jumpers and DIP Switches Jumpers
are small connectors that bridge
across predetermined terminal points
(pins) on the card itself to hardwire
the card for certain user-defined settings, such as the IRQ setting. DIP
(dual inline package) switches are
small switches (usually in groups)
that, like jumpers, can configure the
card for user-defined settings.
Before you buy a network adapter card, you must make sure it has
the correct data bus architecture for your PC and the correct connector type for your transmission medium. It must support the operating system you are running on the computer into which it is being
installed.
Almost all PCs use one of four basic data bus architectures: ISA,
EISA, PCI, and Micro Channel. (Refer to the In-depth on data bus
architectures earlier in this chapter.) These architectures are not necessarily compatible. For example, a Micro Channel card doesn’t work
on an EISA system and, in fact, doesn’t even fit in the slot, so when
you buy a card for an expansion slot, be ready to tell the vendor
what type of data bus architecture you have available on your
system.
Chapter 3, “Transmission Media,” discussed some basic LAN network cabling types. The network adapter is responsible for transmitting in accordance with the specifications of the transmission
medium. The adapter card also must supply a connector that is compatible with the cabling system. (See Chapter 3 for more information on Ethernet and token-ring cabling and connectors.) Some
boards offer connectors for more than one cabling type, in which
case you must configure jumpers or DIP switches or use software to
set the active type.
CONFIGURING NETWORK ADAPTER
CARDS
You must configure your operating system so that it can communicate with the network adapter card. Most plug-and-play adapter
cards configure themselves and, in conjunction with the operating
system running, assign resources to themselves. In many cases,
10.919-x CH05.i
8/28/98 12:56 PM
Page 207
C hapter 5
though, you must manually configure the adapter card. These settings are configured through jumper or DIP switch settings, or by
using some form of software, so that the network card can communicate with the operating system.
To communicate, the operating system and the network adapter
must agree on certain important parameters, called resource settings.
Some common resource settings for a network adapter are as follows:
á IRQ
á Base I/O port address
á Base memory address
á DMA channel
á Boot PROM
á MAC address
á Ring speed (token-ring cards)
á Connector type
IRQ
The IRQ (Interrupt Request Line) setting reserves an interrupt request
line for the adapter to use when contacting the CPU. Devices make
requests to the CPU using a signal called an interrupt. Each device
must send interrupts on a different interrupt request line. Interrupt
request lines are part of the system hardware. The IRQ setting (such
as IRQ3, IRQ5, or IRQ15) defines which interrupt request line the
device is to use. By convention, certain IRQ settings are reserved for
specific devices. Different adapter cards provide numerous available
IRQs from which you can choose. Be careful, though: Some network
cards are very limited in the different IRQs that they have available,
and may offer only IRQs that are currently being used by other
devices on your system.
Base I/O Port Address
The base I/O port address defines a memory address through which
data flows to and from the adapter. The base I/O port address
NETWORK ADAPTER CAR D S
207
10.919-x CH05.i
208
8/28/98 12:56 PM
C h a p te r 5
Page 208
NE TWO R K ADAP TER CARDS
functions more like a port, defining a channel between the adapter
and the processor.
Base Memory Address
The base memory address is a place in the computer’s memory that
marks the beginning of a buffer area reserved for the network
adapter. Not all network adapter cards use the computer’s RAM, and
therefore not all adapters require a base memory address setting.
DMA Channel
The DMA or Direct Memory Access channel is an address used for
quicker access to the CPU by the adapter card. Many devices,
including network cards, often enable you to choose between DMA
channels 1 through 7, or have the channel disabled. Not all network
cards have the capability to set DMA channels.
Boot PROM
Some network adapter cards are equipped with a Boot PROM
(Programmable Read Only Memory). This Boot PROM enables the
network card to boot up and connect over the network, because the
Boot PROM has the necessary connection software to use. This feature is often used by diskless workstations, because they have either
no hard drives or floppy drives onto which to store connection software.
MAC Address
As discussed in Chapter 2, using the MAC address burnt into each
network card is one of the several ways to establish addresses for
nodes on the network. These addresses are hexadecimal in nature
and are unique for each card. The IEEE is responsible for assigning
these addresses to each network card manufacturer. In some cases,
you can reassign a new MAC address for the network adapter card.
In the case of ARCNet cards, DIP switches are used to set the network card’s address.
10.919-x CH05.i
8/28/98 12:56 PM
Page 209
C hapter 5
Ring Speed
In the case of token-ring networks, the ring speed must be set on the
token-ring card. The possible values for this are either 4Mbps or
16Mbps. It is very important that the correct ring speed is set,
because an incorrect ring speed prevents your computer from connecting onto the network, or it also can cause the entire network to
fail.
Connector Type
Some network cards have different connectors from which you can
choose. A common example is an ethernet card with both a BNC
connector and an RJ45 connector. Some network cards require that
the connector to be used must be specified. Other network cards
self-adjust to the connector being used.
Any effort to configure a network adapter card should begin with
the card’s vendor documentation. The documentation tells you
which resource settings are available for you to set, and it might recommend values for some or all of the settings. Some network cards
have a default setting, in which all settable values are set to defaults
recommended by the factory.
The actual process of configuring the operating system to interact
with a network adapter card depends on the operating system. A
plug-and-play operating system such as Windows 95, when used
with a plug-and-play-compatible network adapter card, may perform
much of the configuring automatically. In Windows NT, you can
often configure adapter card resource settings through the Control
Panel Network application’s Adapters tab. (Not all adapter cards are
configurable through this interface.) The Windows NT Diagnostics
application in the Administrative Tools group (see Exercise 5.2 at the
end of the chapter) indicates the resource settings that are currently
available.
RESOLVING HARDWARE CONFLICTS
Hardware conflicts are caused when the devices on the system compete for the same system resources, such as interrupt request lines,
NETWORK ADAPTER CAR D S
209
10.919-x CH05.i
NOTE
210
8/28/98 12:56 PM
C h a p te r 5
Page 210
NE TWO R K ADAP TER CARDS
Plug-and-Play It is important to note
that plug-and-play (sometimes facetiously called plug-and-pray by its critics) is still relatively new technology
for Microsoft-based systems. Ideally,
Windows 95 should configure a plugand-play-compatible card without
much user intervention, but in some
cases, you might still face configuration problems. This occurs because
many plug-and-play devices are simply
not coded correctly to work with plugand-play operating systems.
base I/O port addresses, and base memory addresses. An improperly
configured device can cause a hardware conflict with other devices,
so you must make sure that each device has exclusive access to the
required system resources. An example of common resources and the
devices assigned to them is seen in Table 5.1.
Other devices installed into a computer that often take up resources
are sound cards, modems (discussed in Chapter 6), scanners, and
CD-ROM controllers.
In Windows NT, a hardware conflict might evoke a warning message
from the system or an entry in the Event Log (see Chapter 12,
“Troubleshooting”). If you experience a hardware conflict, use
Windows NT Diagnostics (see Exercise 5.2) to check resource settings for system devices. Then change the resource settings of any
conflicting devices.
In Windows 95, use Device Manager (see the following Step by
Step) to spot hardware conflicts and track resource settings.
TABLE 5.1
C O M M O N R E S O U R C E A L L O C AT I O N
Resources
IRQ
System Timer
0
Key Board
1
060-06F
VGA display
2/9
3C0-3CF
COM2
3
2F8-2FF
COM1
4
3F8-3FF
LPT2
5
278-27F
Floppy
I/O Address
3F0-3F7
Controller
LPT1
7
378-37F
Math coprocessor
13
0F0-0F8
Primary hard
drive controller
14
1f0-1f8
Secondary hard
drive controller
15
170-177
ON A
PC
Memory
DMA
A0000-AFFFF
0
10.919-x CH05.i
8/28/98 12:56 PM
Page 211
C hapter 5
NETWORK ADAPTER CAR D S
STEP BY STEP
5.2 Using Device Manager
Windows 95 includes a utility called Device Manager that displays
system devices by type, looks for resource conflicts, and provides an
interface for checking and changing resource settings.
To access Device Manager, follow these steps:
1. Click the Start button and choose Settings, Control Panel.
2. In the Windows 95 Control Panel, double-click the
System application.
3. Choose the Device Manager tab in the System Properties
dialog box.
4. Device Manager displays system devices in a tree format.
Click on the plus sign next to a device type to view the
installed devices. Double-click on an installed device (or
choose the device and click the Properties button) for a
Properties dialog box, such as the one shown in Figure
5.11.
Also, by selecting the computer icon and clicking the
properties button, you see a list of all resource usage for
Memory addresses, IRQs, DMA channels, and
Input/Output settings.
If you can’t pinpoint a resource conflict by using Windows NT
Diagnostics, Windows 95’s Device Manager, or some other diagnostic program, try removing all the cards except the network adapter
and then replacing the cards one by one. Check the network with
each addition to determine which device is causing the conflict.
FIGURE 5.11
An adapter card’s Properties dialog box in
Device Manager.
211
10.919-x CH05.i
212
8/28/98 12:56 PM
C h a p te r 5
Page 212
NE TWO R K ADAP TER CARDS
C A S E S T U DY : W O R K I N G
E S TA B L I S H E D N E T W O R K
WITH
ADAPTER CARDS
ESSENCE OF THE CASE
The facts for this case study are as follows:
• You have no documentation for the following:
• What hardware is installed on these
computers
• What software is installed on the systems
• How the network is configured
• You have no software for these cards;
this includes configuration software or
the drivers themselves.
• Based upon the description of the network adapter cards, you know that this
network is a token-ring network.
• The network card manufacturer is Intel.
SCENARIO
You have been contacted by a firm to come in
and attach all 10 of their new PCs to their existing token-ring network. However, the network
adapter cards they want to use have been
stripped out of old computers and are being
reused in the new PCs. You try to obtain the
detailed records of the network, but find that no
one has been keeping records, thus you are
going to have to obtain all information to install
and configure these network adapter cards on
your own. The network adapter cards are software-configurable, but a thorough search has
failed to turn up the software needed to configure
IN AN
these network cards. You can see from the writing on the network adapter card that it is an Intel
Token Express 16/4. What are the steps you
would take to attach these computers to the network?
A N A LY S I S
This case study is an example of a problem that
almost all network professionals have encountered in their lives. This problem is made extra
unique, for one issue is that you must not only
resolve any hardware, or possible hardware conflicts, but you must also somehow obtain the
software to configure these token-ring cards.
The approach to take in this situation is as follows: First, you need to acquire the software to
configure these cards. Often network adapter
card manufactures place the configuration software and drivers up on the web. But this is not
always the case. Numerous network card manufacturers make “compatible” network cards, yet
the software does not always work with the cards
with which they are reported to be compatible.
Another issue with network adapter cards is that
in some cases no identifiable markings are on
the network adapter card to enable you to even
identify the manufacturer or model of the card. In
this case study you are at least able to determine that this card is an Intel Token Express
16/4.
In this case, if you go to Intel’s Web site and do
a search for the keywords of “Intel Token Express
16/4,” you get a list back that includes a downloadable compressed file containing the drivers
and configuration software for this card.
10.919-x CH05.i
8/28/98 12:56 PM
Page 213
C hapter 5
C A S E S T U DY : W O R K I N G
E S TA B L I S H E D N E T W O R K
WITH
NETWORK ADAPTER CAR D S
ADAPTER CARDS
After obtaining this compressed file, you should
uncompress it. The first thing you should look
for is whether any information files are available
that may note installation issues concerning the
network adapter card. These issues can range
from what hardware is incompatible with this
network adapter card to what software may not
work with these cards. This information is often
found in the README.TXT or READ.ME file or
may be part of the configuration software. To
resolve these conflicts if they exist, you should
check whether any of the PCs in which you are
installing these network adapter cards contain
this problematic software or hardware. If these
213
IN AN
components exist, you should either replace
them or find a possible software fix.
The next step in this installation process, assuming the machines are running Windows 95 or
Windows NT, is to go to either the Device manager (Windows 95) or the Windows NT Diagnostics
(Windows NT) to determine what resources (that
is, IRQ, DMA, I/O, and so forth) are not being
used. Based upon this information, the proper
values for the network card can be chosen.
Because this network is a token-ring network,
you should check to make sure that you select
the correct ring speed at which the card should
operate.
CHAPTER SUMMARY
This chapter examined the network adapter card—an essential component in ethernet and token-ring networks. The network adapter
card performs several functions, including preparing, sending, and
controlling the flow of data to the network transmission medium.
To address these issues, this chapter focused on the following main
issues:
KEY TERMS
• Data bus
• Analog signal
• Digital signal
• Clocking
á Defining the workings of a network adapter card
• Current state
á How to install a network card
• State transmission
á How to configure a network adapter card
á How to resolve hardware conflicts
When you prepare for the exam, pay particular attention to the last
three points, because it is on these issues that the exam objective
focuses. The first point was included to give you needed background
and a better understanding of the functionality of a network adapter
card.
• IRQ
• Base I/O port address
• Base memory address
• DMA channel
• Boot PROM
• MAC address
• Ring speed
• Connector type
10.919-x CH05.i
214
8/28/98 12:56 PM
C h a p te r 5
Page 214
NE TWO R K ADAP TER CARDS
A P P LY Y O U R L E A R N I N G
Exercises
5.1
Network Adapter Resource Settings
Objective: Become familiar with the process of
configuring network adapter resource settings in
Windows NT.
Estimated time: 10 minutes
Earlier in this chapter, you learned how to install a network adapter card driver by using the Windows NT
Network application. You can also use the Network
application to check or change the resource settings for
an adapter that is already installed.
1. Click the Start button and choose
Settings/Control Panel. Double-click the
Windows NT Control Panel Network application.
2. In the Network application, click the Adapters
tab.
3. Select the network adapter that is currently
installed on your system and click the Properties
button.
4. The Network Card Setup dialog box then appears
on your screen (see Figure 5.12).
5. In the Network Card Setup dialog box, you can
change the resource settings as required. You
might want to use the Windows NT Diagnostics
application to look for available settings (see
Exercise 5.2). Don’t change the settings unless
you’re experiencing problems, though, because
you could introduce a hardware conflict with
another device.
6. Click Cancel to leave the Network Card Setup
dialog box and click Cancel again to leave the
Network application.
FIGURE 5.12
A Network Card Setup dialog box.
5.2
Windows NT Diagnostics
Objective: Learn to check resource settings through
Windows NT Diagnostics.
Estimated time: 10 minutes
Windows NT Diagnostics tabulates a number of
important system parameters. You can use Windows
NT Diagnostics to help resolve resource conflicts for
network adapters.
1. Click the Start button and choose
Programs/Administrative Tools. Choose Windows
NT Diagnostics from the Administrative Tools
menu.
2. Windows NT Diagnostics provides several tabs
with information on different aspects of the system. Choose the Resources tab (see Figure 5.13).
10.919-x CH05.i
8/28/98 12:56 PM
Page 215
C hapter 5
NETWORK ADAPTER CAR D S
215
A P P LY Y O U R L E A R N I N G
FIGURE 5.13
FIGURE 5.14
The Windows NT Diagnostics Resource tab.
The Windows NT Diagnostics Resource tab—
I/O port setting
3. Figure 5.13 displays the IRQ settings for system
devices. (Note that the network adapter card for
which the resource settings were displayed is listed here beside IRQ2.) The buttons at the bottom
of the screen invoke views of other resource settings. Click on a button to see the associated list.
Figure 5.14 shows the I/O Port list.
You can’t change any values in Windows NT
Diagnostics. You can only view services, devices, statistics, and settings. Changes to devices can be done from
their respective control panel applets. Network cards
are often configurable from the Network applet and
SCSI devices are changed under the SCSI applet, to
name two examples. However, not all devices can be
configured from their respective applets. Some require
you to use manufacturer-supplied software or even
adjust jumper settings to make any configuration
changes.
Review Questions
1. What type of signal has a discrete state?
2. What are the configurable options on a network
card?
3. What can be set on a token-ring card but not on
an ethernet card?
4. What is caused by an incorrect setting on a network card?
Exam Questions
1. In Windows NT, what is a utility you can use to
install network adapter card drivers?
A. Windows NT Diagnostics
10.919-x CH05.i
216
8/28/98 12:56 PM
C h a p te r 5
Page 216
NE TWO R K ADAP TER CARDS
A P P LY Y O U R L E A R N I N G
B. The System application
C. Device Manager
D. None of the above
6. Which resource setting specifies a serial communications port for the network adapter?
A. IRQ
B. Base I/O port address
2. What must a user sometimes configure to hardwire resource settings on a network adapter card?
A. Jumpers
B. Resource switches
C. Needle connectors
D. None of the above
C. Base memory address
D. None of the above
7. Which resource setting locates a buffer for the
adapter in the computer’s RAM?
A. IRQ
B. Base I/O port address
3. Which two of the following are common data
bus architectures?
A. EISA
B. Pentium
C. Plug-and-play
D. PCI
4. Which resource setting gives the device a channel
for contacting the CPU?
A. IRQ
C. Base memory address
D. None of the above
8. Which two of the following enable you to check
the resource settings for a network adapter card in
Windows NT?
A. Device Manager
B. The Network application
C. Windows NT Diagnostics
D. The System application
B. Base I/O port address
C. Base memory address
D. None of the above
5. Which resource setting defines a means for passing data to the adapter?
9. Which of the following enables you to change the
resource settings for a network adapter card in
Windows NT?
A. Device Manager
B. The Network applet
A. IRQ
C. Windows NT Diagnostics
B. Base I/O port address
D. The System applet
C. Base memory address
D. None of the above
10.919-x CH05.i
8/28/98 12:56 PM
Page 217
C hapter 5
NETWORK ADAPTER CAR D S
217
A P P LY Y O U R L E A R N I N G
10. What are three duties of the network adapter
card?
A. Preparing data
B. Sending data
C. Identifying problems with the cabling
medium
D. Controlling the flow of data
11. On ethernet networks, data flows from the network adapter card to the transmission medium in
what form?
A. Parallel
B. Serial
C. Either A or B
D. None of the above
12. Of the following IRQs, which would be a recommended IRQ setting for a network adapter card?
A. IRQ14
B. IRQ2
C. IRQ1
D. IRQ5
13. A user complains that after having a network card
installed in his system, he is knocked off the network whenever he prints. What is the most likely
solution to this problem?
A. The printer’s cable is plugged into the network card.
B. The newly installed network card’s IRQ is in
conflict with the printer port’s IRQ.
C. The newly installed network card is not seated correctly.
D. The newly installed network card’s IRQ is in
conflict with the sound card’s IRQ.
14. A computer’s network adapter card is not functioning. You suspect a resource conflict with
another device. The operating system being used
by the computer is Windows 95.
Primary Objective: Make the network adapter
card work.
Secondary Objective: Learn what resources are
being used by the network adapter card.
Secondary Objective: Change the resource settings on the network adapter card.
Suggested Solution: You go into Windows NT
Diagnostics and view the resource configuration
information. Based upon this information, you
reset the network adapter card’s resources in the
Network section of Windows NT diagnostics.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objective.
D. This solution does not meet the primary
objective.
Answers to Review Questions
1. Digital signals have discrete states. These two discrete states are “on” and “off.” Analog signals follow a wave pattern that has both amplitude and
frequency. See the section titled “Signals.”
10.919-x CH05.i
218
8/28/98 12:56 PM
C h a p te r 5
Page 218
NE TWO R K ADAP TER CARDS
A P P LY Y O U R L E A R N I N G
2. Most common configurable options on a network
card are:
• IRQ
all possibilities stated in the second scenario. See
the section titled “Resolving Hardware
Conflicts.”
• I/O port
• Memory address
• DMA
• MAC address
• Connector type
• Ring speed (token ring)
See the section titled “Configuring Network
Adapter Cards.”
3. A token-ring card can set the ring speed, which is
not an option on an ethernet card. Some ethernet
cards can set their data transfer rates between
10Mbps and 100Mbps, but this is not the same
as ring speed. The options for ring speed are
4Mbps and 16Mbps. Incorrect setting of this can
cause the card to be inoperable or can shut down
the entire network. See the section titled “Ring
Speed.”
4. An incorrect setting on a network card causes one
of several things to occur:
First: An incorrect setting can cause the network
adapter card to not function.
Second: An incorrect setting can cause another
device on the computer to not function, because
this other device may also be using the resource
set on the network adapter card. As this happens,
the network adapter may or may not also function.
Third: An incorrect setting can cause another
device on the network to not function, as well as
Answers to Exam Questions
1. D. The correct answer would be the Network
icon in Control panel. B is not used to configure
network adapters. C can be used in Windows 95.
See the section titled “Installing Network Adapter
Cards.”
2. A. Jumpers are used to configure resource settings
on network adapter cards. Other valid options
could also have been DIP switches and software
configuration utilities. See the section titled
“Configuring Network Adapter Cards.”
3. A, D. B is a processor type; C is an architecture
standard to which devices can conform. See the
section titled “Defining the Workings of a
Network Adapter Card.”
4. B. A is used to signal the processor that a device
is ready to use the processor. C is a storage area
for storing data that is passing to and from the
device. See the section titled “Configuring
Network Adapter Cards.”
5. B. A is used to signal the processor that a device
is ready to use the processor. Data is passed along
the Base I/O port address. C is a storage area for
storing data that is passing to and from the
device. See the section titled “Configuring
Network Adapter Cards.”
6. D. A serial communication port is specified as a
COMx port, where x typically ranges from 1-4.
See the section titled “Configuring Network
Adapter Cards.”
10.919-x CH05.i
8/28/98 12:56 PM
Page 219
C hapter 5
NETWORK ADAPTER CAR D S
219
A P P LY Y O U R L E A R N I N G
7. C. A is used to signal the processor that a device
is ready to use the processor. C is used for buffering data on a computer’s RAM. See the section
titled “Configuring Network Adapter Cards.”
8. B, C. Using the Device Manager and the system
application (where Device Manager is actually
found) are methods of checking for resource settings in Windows 95. See the section titled
“Resolving Hardware Conflicts.”
9. B. Only through the network application can
resource changes be made. All other answers are
read-only utilities found in Windows 95 and NT.
It should be noted that some devices do not allow
changes to be made through the Network
Application. See the section titled “Resolving
Hardware Conflicts.”
10. A, B, D. Problems with the cable medium are
addressed with specific diagnostic utilities that are
discussed in Chapter 11, “Monitoring the
Network.” See the section titled “Defining a
Network Adapter.”
11. B. The bits are sent out one after the other, just
as they are in a com port. In parallel form bits are
sent out eight at a time. See the section titled
“Preparing Data.”
12. D. IRQ 1 is used by the Keyboard; IRQ2 is used
by the display adapter; IRQ14 is used by the
Primary Hard drive Controller. See the section
titled “Configuring Network Adapter Cards.”
13. B. If a network card’s IRQ is set to the same one
used by the printer, the printer can cause the
adapter card to not function anymore, thus causing a user’s computer to be knocked off the network. A would not even enable the network card
or printer to work in the first place. C would
cause the network card to not function at all.
Answer D would cause conflicts between the
sound card and Network adapter card. See the
section titled “Configuring Network Adapter
Cards.”
14. D. NT Diagnostics is a utility that enables one to
view only configuration information. Nothing
can be set through this utility. Also, there could
be many reasons why the network adapter card is
not functioning. It could be broken, unplugged
from the network, incompatible with the operating system, or have a resource conflict with
another device. The proposed solution satisfies
only the first secondary objective. See the section
titled “Resolving Hardware Conflicts.”
Suggested Readings and Resources
1. Anderson, Douglas T. The network interface
card technical guide. Micro House, 1993.
2. Derfler, Frank J., Jr., Using Networks. Que,
1998.
• Chapter 5: LAN Adapters: The Hardware
Heart of the LAN.
10.919-x CH05.i
8/28/98 12:56 PM
Page 220
11.919-x CH06.i
8/28/98 1:03 PM
Page 221
OBJECTIVES
Chapter 6 targets the following objective in the
Planning section of the Networking Essentials exam:
Select the appropriate connectivity devices for
various token--ring and Ethernet networks.
Connectivity devices include repeaters, bridges,
routers, brouters, and gateways
. This exam topic was briefly highlighted in Chapter
2, “Networking Standards.” In this chapter, a more
in-depth analysis is done on the different devices
that are used on various token-ring and Ethernet
networks. In this chapter, besides simply addressing
the role of each connectivity device, emphasis is
placed on comparing and contrasting the devices in
terms of when one would be used versus another.
C H A P T E R
6
Connectivity Devices
and Transfer
Mechanisms
11.919-x CH06.i
8/28/98 1:03 PM
Page 222
OUTLINE
S T U DY S T R AT E G I E S
Addressing
223
Modems
224
Asynchronous Transmission
225
Synchronous Transmission
227
Repeaters
228
Hubs
230
Passive Hubs
231
Active Hubs
232
Intelligent Hubs
232
Bridges
233
Routing
235
Routers
236
Brouters
239
Gateways
239
Dynamic Routing Applied–Routing
Algorithms
240
Distance Vector Routing
241
Link-State Routing
242
Chapter Summary
244
. When reading this chapter be very aware of the
differences that each connectivity device plays.
Some connectivity devices can perform multiple
roles. If you can address when and why a certain connectivity device needs to be used, you
will have no problems addressing the exam
topic being presented by this chapter.
11.919-x CH06.i
8/28/98 1:03 PM
Page 223
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
223
INTRODUCTION
People sometimes think of a network as a single, local cabling system that enables any device on the network to communicate directly
with any other device on the same network. A network by this definition, however, has no connections to other remote networks.
An internetwork consists of multiple independent networks that are
connected and can share remote resources. These logically separate
but physically connected networks can be dissimilar in physical type
and topology. The device that connects the independent networks
together may need a degree of “intelligence” because it may need to
determine when packets will stay on the local network or when they
will be forwarded to a remote network.
This chapter examines some important connectivity devices. In
order to facilitate your understanding of what these devices actually
do, the chapter begins by explaining the concept of addressing. This
is an important concept to understand in order to differentiate
between different connectivity devices that exist on the network. In
the following sections, you learn about modems, repeaters, bridges,
routers, brouters, and gateways. Some of this material also appears
in Chapter 2, “Networking Standards,” in the discussion of communication devices and OSI. In Chapter 2, the main emphasis was
where the different connectivity devices were situated within the
OSI model. Here the emphasis is more on the function of the different connectivity devices, and when one would use them. The chapter concludes with a section discussing different routing algorithms.
ADDRESSING
Before a discussion of devices is warranted, addressing of a network
must be explored further. Addressing a network is important,
because it is by this mechanism that devices on the network are
located and identified.
A network is similar to a city. A city has buildings and it has streets.
Now imagine a city in which no two streets had the same name.
Each street name must be unique. Now imagine a city that not only
had unique names for each street, but also a unique address for each
building on its street (this should not be too difficult).
<$Inetworks;addressing;overvie
w><$Iaddressing;networks;overv
iew>
11.919-x CH06.i
224
8/28/98 1:03 PM
C h a p te r 6
Page 224
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
<$Inetworks;addressing;ov
erview><$Iaddressing;networks;overview>
A network is like a street. All buildings on a single street share the
same street name that they reside on. Or in other words, all devices
on a network segment share the same network address. Thus there
are two distinct forms of addresses here: the logical network address
and the node address that is physically part of the device.
All packets that go on to a network have within them the source
address information and destination address information. Some
packets are not routable; therefore they will not contain a source or
destination network address in them. The details of routing will be
covered in Chapter 7, “Transport Protocols.”
These concepts have been touched on in previous chapters, but it is
important that you understand these two concepts here, as the function of different devices on the network is dependent upon device
(or hardware) addresses and logical network addresses. As seen in the
previous chapter, device addresses in terms of a network card were
either burnt onto, or programmed into, the network card by the
manufacturer. These addresses could also be set with software by
some network card manufacturers, or could be set through DIP
switches, as in the case of ARCNet cards. This chapter will address
different routing protocols used to discover the existence of different
logical addresses on the network, whereas the next chapter will
address rules that different transport protocols conform to when
addressing networks.
MODEMS
<$Imodems;terminology><$Imodems;signal
conversion
function><$Imodems;line
Standard telephone lines can transmit only analog signals.
Computers, however, store and transmit data digitally. Modems can
transmit digital computer signals over telephone lines by converting
them to analog form.
Converting one signal form to another (digital to analog in this case)
is called modulation. Recovering the original signal is called demodulation. The word “modem” derives from the terms
modulation/demodulation.
Modems can be used to connect computer devices or entire networks that are at distant locations. (Before digital telephone lines
existed, modems were about the only way to link distant devices.)
11.919-x CH06.i
8/28/98 1:03 PM
Page 225
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
225
Modems enable networks to exchange email and to perform limited
data transfers, but the connectivity made possible is extremely limited due to the limited bandwidth most modems offer. Modems don’t
enable networks to connect to remote networks, like a router, to
directly exchange data. Instead modems act like network cards in
that they provide an access point onto the transmission medium, in
this case the telephone lines, in order to send analog signals to
another device, most likely another modem, on the network.
Until recently, modem manufacturers used a parameter called baud
rate to gauge modem performance. The baud rate is the oscillation
speed of the sound wave transmitted or received by the modem.
Although baud rate is still an important parameter, recent advances
in compression technology have made it less meaningful. Some
modems now provide a data transfer rate (in bits per second—a
more meaningful measure of network performance) that exceeds the
baud rate. In other words, you can no longer assume the baud rate
and the data transfer rate are equal.
NOTE
Some modems operate constantly over dedicated phone lines.
Others use standard public switched-telephone network (PSTN)
dial-up lines and make a connection only when one is required.
Other Modem Uses Modems don’t
necessarily need to connect through
the PSTN. Short-haul modems frequently are used to connect devices
in the same building. A standard serial connection is limited to 50 feet, but
short-haul modems can be used to
extend the range of a serial connection to any required distance.
Many devices are designed to operate
with modems. When you want to connect such devices without using
modems, you can use a null-modem
cable, which connects the transmitter
of one device to the receiver of the
other device.
Modems are classified according to the transmission method they
use for sending and receiving data. The two basic types of modems
are as follows:
á Asynchronous modems
á Synchronous modems
The following sections describe asynchronous and synchronous
transmission.
Asynchronous Transmission
Asynchronous transmission does not use a clocking mechanism to
keep the sending and receiving devices synchronized. Instead, this
type of transmission uses bit synchronization to synchronize the
devices for each frame that is transmitted.
In bit synchronization, each frame begins with a start bit that
enables the receiving device to adjust to the timing of the
<$Imodems;line
types;dedicated
><$Imodems;line
types;dial-up
><$Imodems;baud
rates><$Imodems;ty
pes;asynchronouns><$Imodems;t
11.919-x CH06.i
226
8/28/98 1:03 PM
C h a p te r 6
Page 226
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
<$Iasynchronous
modems;bit synchronization><$Imodems;types;a
synchronous><$Iasynchr
onous modems;frame
transmissions><$Iasynch
ronous modems;high
overhead><$Iasynchrono
us
modems;uses><$Iasynch
ronous modems;hardware costs>
transmitted signal. Messages are kept short so that the sending and
receiving devices do not drift out of synchronization for the duration
of the message. Asynchronous transmission is most frequently used
to transmit character data and is ideally suited to environments in
which characters are transmitted at irregular intervals, such as when
users enter character data.
Figure 6.1 illustrates the structure of a typical frame used to transmit
character data. This frame has four components:
á A Start bit. This component signals that a frame is starting and
enables the receiving device to synchronize itself with the message.
á Data bits. This component consists of a group of seven or
eight bits when character data is being transmitted.
á A parity bit. This component is optionally used as a crude
method of detecting transmission errors.
á A stop bit or bits. This component signals the end of the data
frame.
Asynchronous transmission is a simple, inexpensive technology ideally suited for transmitting small frames at irregular intervals. Because
start, stop, and parity bits must be added to each character being
transmitted, however, overhead for asynchronous transmission is
high—often in the neighborhood of nearly 20 to 30 percent. This
high overhead wastes bandwidth and makes asynchronous transmission undesirable for transmitting large amounts of data.
Asynchronous transmission is frequently used for PC-to-PC and
terminal-to-host communication. Modems use asynchronous transmission. Data in these environments is often of the bursty, characteroriented nature that is ideal for asynchronous communication.
Asynchronous transmission generally requires less expensive hardware than synchronous transmission.
FIGURE 6.1
The structure of an asynchronous frame consists of four key bit components.
Start
Bit
(1)
Data
Bits
(7-8)
Parity
Bit
(0-1)
Stop
Bits
(1-2)
11.919-x CH06.i
8/28/98 1:03 PM
Page 227
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
Synchronous Transmission
Synchronous transmission eliminates the need for start and stop bits
by synchronizing the clocks on the transmitting and receiving
devices. This synchronization is accomplished in two ways:
á By transmitting synchronization signals with data. Some data
encoding techniques, by guaranteeing a signal transition with
each bit transmitted, are inherently self-clocking.
á By using a separate communication channel to carry clock sig-
nals. This technique can function with any signal-encoding
technique.
Figure 6.2 illustrates the two possible structures of messages associated with synchronous transmission.
<$Imodems;types;synchronous><$Isynchronous
modems;message structures><$Isynchronous
modems;data
types><$Isynchronous
modems;cyclic redundancy
checks (CRC )><$Icyclic
redundancy check (CRC)>
Both synchronous transmission methods begin with a series of synch
signals, which notify the receiver of the beginning of a frame. Synch
signals generally utilize a bit pattern that cannot appear elsewhere in
messages, ensuring that the signals always are distinct and easily recognizable by the receiver.
A wide variety of data types can be transmitted. Figure 6.2 illustrates
both character-oriented and bit-oriented data. Notice that under
synchronous transmission, multiple characters or long series of bits
can be transmitted in a single data frame. Because the transmitter
and receiver remain in synchronization for the duration of the transmission, frames may be very long.
When frames are long, parity is no longer a suitable method for
detecting errors. If errors occur, multiple bits are more likely to be
affected, and parity techniques are less likely to report an error. A
more appropriate error-control technique for synchronous transmission is the cyclic redundancy check (CRC). In this technique, the
SYNCH
SYNCH
CHARACTER
SYNCH
SYNCH
BINARY DATA
• • •
CRC
CHARACTER
END
CRC
FILL BITS
END
SYNCH
SYNCH
DATA
CRC
END
FIGURE 6.2
Structures of synchronous transmission.
227
11.919-x CH06.i
228
8/28/98 1:03 PM
C h a p te r 6
Page 228
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
<$Isynchronous
modems;versus asynchronous
modems><$Iasynchron
ous modems;versus synchronous modems>
transmitter uses an algorithm to calculate a CRC value that summarizes the entire value of the data bits. This value is then appended to
the data frame. The receiver uses the same algorithm, recalculates the
CRC, and compares the CRC in the frame to the CRC value it has
calculated. If the values match, the frame almost definitely was transmitted without error.
When synchronous transmission links are idle, communicating
devices generally send fill bits to the devices synchronized.
Synchronous transmission offers many advantages over asynchronous
transmission. The overhead bits (synch, CRC, and end) comprise a
smaller portion of the overall data frame, which provides for more
efficient use of available bandwidth. Synchronization improves error
detection and enables the devices to operate at higher speeds.
The disadvantage of synchronous transmission is that the more complex circuitry necessary for synchronous communication is more
expensive. Network adapter cards commonly employ synchronous
transmission methods.
REPEATERS
<$Irepeaters;defined><$It
ransmission
media;repeaters><$Irepea
ters;signal
generation><$Irepeaters;c
able
ranges><$Icables;repeater
use><$Irepeaters;inex-
As you learned in Chapter 3, “Transmission Media,” all media attenuate the signals they carry. Each media type, therefore, has a maximum range that it can reliably carry data. The purpose of a repeater
is to extend the maximum range for the network cabling.
A repeater is a network device that repeats a signal from one port
onto the other ports to which it is connected (see Figure 6.3).
Repeaters operate at the OSI Physical layer. (Refer to “The OSI
Reference Model” section in Chapter 2.) A repeater does not filter or
interpret—it merely repeats (regenerates) a signal, passing all network traffic in all directions.
A repeater doesn’t require any addressing information from the data
frame because a repeater merely repeats bits of data. This means that
if data is corrupt, a repeater will regenerate the signal anyway. A
repeater will even repeat a broadcast storm caused by a malfunctioning adapter (see Chapter 12, “Troubleshooting”).
The advantages of repeaters are that they are inexpensive and simple.
Also, although they cannot connect networks with dissimilar data
11.919-x CH06.i
8/28/98 1:03 PM
Page 229
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
229
FIGURE 6.3
Repeater
A repeater regenerates a weak signal.
Strong
Signal
(Regenerated)
Weak Signal
frames (such as a token-ring network and an Ethernet network),
some repeaters can connect segments with similar frame types but
dissimilar cabling.
Figure 6.4 shows the use of a repeater to connect two Ethernet cable
segments. The result of adding the repeater is that the potential
length of the overall network is doubled.
<$Irepeaters;Ethernet cable
diagram><$Irepeaters;ampli
fication
feature><$Irepeaters;length
limitations>
Some repeaters simply amplify signals. Although this increases the
strength of the data signal, it also amplifies any noise on the network. In addition, if the original signal has been distorted in any
way, an amplifying repeater cannot clean up the distortion.
Certainly, it would be nice if repeaters could be used to extend
networks indefinitely, but all network designs limit the size of the
network. The most important reason for this limitation is signal
propagation. Networks must work with reasonable expectations
DIRECT
ATTACHMENT
TO T CONNECTOR
TERMINATOR
GROUNDED
TERMINATOR
REPEATER
MINIMUM DISTANCE
.5 METER
185 METERS MAXIMUM
UP TO 30 ATTACHMENTS
FIGURE 6.4
Using a repeater to extend an Ethernet LAN.
11.919-x CH06.i
230
8/28/98 1:03 PM
C h a p te r 6
Page 230
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
<$Irepeaters;propagation
delay><$Ipropagation
delay;repeaters><$Irepeaters;5-4-3
rule><$I5-4-3 rule (repeaters)>
about the maximum time a signal might be in transit. This is known
as propagation delay—the time it takes for a signal to reach the farthest point on the network. If this maximum propagation delay
interval expires and no signals are encountered, a network error condition is assumed. Given the maximum propagation delay allowed, it
is possible to calculate the maximum permissible cable length for the
network. Even though repeaters enable signals to travel farther, the
maximum propagation delay still sets a limit to the maximum size of
the network. Repeaters also follow the 5-4-3 rule. That is five segments connected by four repeaters, with no more than 3 of the segments being populated.
NOTE
HUBS
Hub vs. MAU Token-ring networks
use Multiple Access Units (MAUs).
Although they perform a similar function to a hub, they are not interchangeable with a hub. That is a hub
cannot be used in place of a MAU,
and a MAU cannot be used in place
of a hub.
Hubs, also called wiring concentrators, provide a central attachment
point for network cabling (see Figure 6.5). Hubs come in three
types:
á Passive
á Active
á Switching
The following sections describe each of these types in more detail.
<$Ihubs;types;passive><$Ihubs;
types;active><$Ihubs;types;swit
ching><$Iwiring concentrators>
FIGURE 6.5
A network wired to a central hub.
11.919-x CH06.i
8/28/98 1:04 PM
Page 231
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
231
Passive Hubs
Passive hubs do not contain any electronic components and do not
process the data signal in any way. The only purpose of a passive
hub is to combine the signals from several network cable segments.
All devices attached to a passive hub receive all the packets that pass
through the hub.
Because the hub doesn’t clean up or amplify the signals (in fact, the
hub absorbs a small part of the signal), the distance between a computer and the hub can be no more than half the maximum permissible distance between two computers on the network. For example, if
the network design limits the distance between two computers to
200 meters, the maximum distance between a computer and the
hub is 100 meters.
<$Ihubs;types
;passive><$Ipa
ssive
hubs;purpose><$Ipassi
ve hubs;limited functionality><$Ipassive
hubs;5-4-3
rule><$I5-4-3
rule;passive
As you might guess, the limited functionality of passive hubs makes
them inexpensive and easy to configure. That limited functionality,
however, is also the biggest disadvantage of passive hubs. Often
small networks use passive hubs, due to the fact that there are few
machines on the LAN and small distances between them. Hubs follow the 5-4-3 rule, where no more than 5 network segments can be
connected to 4 hubs with only 3 of them being populated. This is
seen in Figure 6.6.
FIGURE 6.6
5-4-3 rule: 5 segments connected with 4 hubs,
and only 3 segments are populated.
11.919-x CH06.i
232
8/28/98 1:04 PM
C h a p te r 6
Page 232
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
Active Hubs
<$Ihubs;types;active><
$Iactive hubs;signal
regeneration><$Imulti
port
repeaters><$Isignal
regeneration>
Active hubs incorporate electronic components that can amplify and
clean up the electronic signals that flow between devices on the network. This process of cleaning up the signals is called signal regeneration. Signal regeneration has the following benefits:
á The network is more robust (less sensitive to errors).
á Distances between devices can be increased.
These advantages generally outweigh the fact that active hubs cost
considerably more than passive hubs.
Earlier in this chapter, you learned about repeaters, devices that
amplify and regenerate network signals. Because active hubs function
in part as repeaters, they occasionally are called multiport repeaters.
Intelligent Hubs
Intelligent hubs are enhanced active hubs. Several functions can add
intelligence to a hub:
<$Ihubs;types;intelligent><$Iintelligent
hubs;management
function><$Iintelligent
hubs;switching function>
á Hub management. Hubs now support network management
protocols that enable the hub to send packets to a central network console. These protocols also enable the console to control the hub; for example, a network administrator can order
the hub to shut down a connection that is generating network
errors.
á Switching. The latest development in hubs is the switching
hub, which includes circuitry that very quickly routes signals
between ports on the hub. Instead of repeating a packet to all
ports on the hub, a switching hub repeats a packet only to the
port that connects to the destination computer for the packet.
Many switching hubs have the capability of switching packets
to the fastest of several alternative paths. Switching hubs are
replacing bridges and routers on many networks.
In essence, a switching hub acts like a very fast bridge, which is what
is described in the next section. Switching hubs are the most expensive of hubs on the market. Often they are simply referred to as
Switches. As for the exam, think of all types of hubs as simply a hub.
11.919-x CH06.i
8/28/98 1:04 PM
Page 233
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
BRIDGES
<$Ibridges;signal separation
diagram><$Ibridges;packet
Each packet that is transmitted bears the address of the device to
which it should be delivered. The process works as follows (refer to
Figure 6.7):
NOTE
Bridges, on the other hand, can extend the maximum size of a network. Although the bridged network in Figure 6.7 looks much like
the earlier example of a network with a repeater, the bridge is a
much more flexible device. Bridges operate at the MAC sublayer of
the OSI Data Link layer (see Chapter 2).
A repeater passes on all signals that it receives. A bridge, on the
other hand, is more selective and passes only those signals targeted
for a computer on the other side. A bridge can make this determination because each device on the network is identified by a unique
physical address.
233
Physical Addresses of Network
Adapter Cards Remember that the
physical address of network adapter
cards are often burnt onto the card,
where as in the case of ARCNet, it is
set by DIP switches. See Chapter 5
for more details on this issue.
1. The bridge receives every packet from either side of it on
LAN A.
Node=B
Node=A
Node=B
Node=B
Node=B
LAN A —Side 1
Bridge
Node=A
Node=A
LAN A —Side 2
FIGURE 6.7
Node=A
Separating signals on a LAN segment with a
bridge.
11.919-x CH06.i
234
8/28/98 1:04 PM
C h a p te r 6
Page 234
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
<$Ibridges;packet
transmission
process><$Ibridges
;types;transparent><$Ibridges;typ
es;source routing><$Itransparent
bridges><$Isource
routing
bridges><$Ibridges
;network segmentation>
2. The bridge references an internal table of addresses. This table
is either learned by the bridge, from previous packet deliveries
on the network, or manually programmed into the bridge.
3. Packets on LAN A – Side 1 that are addressed to devices on
LAN A – Side 1 and packets on LAN A – Side 2 that are
addressed to devices on LAN A – Side 2, are not passed along
to the other side by the bridge. These packets can be delivered
without the help of the bridge.
4. Packets on LAN A – Side 1 addressed to devices on LAN A –
Side 2 are retransmitted, by the bridge to LAN A – Side 2 for
delivery. Similarly, the appropriate packets on LAN A – Side 2
are retransmitted to LAN A – Side 1.
Bridges come in two main forms. One type of bridge is what is
known as a transparent or learning bridge. This type of bridge is
transparent to the device sending the packet. At the same time this
bridge will learn over time what devices exist on each side of it. This
is done by the bridge’s ability to read the Data-Link information on
each packet going across the network. By analyzing these packets,
and seeing the source MAC address of each device, the bridge is able
to build a table of which devices exist on what side of it. There usually is a mechanism for a person to go in and also program the
bridge with address information as well. Learning bridges function as
described in step 2, automatically updating their address tables as
devices are added to or removed from the network. Ethernet networks almost always use a transparent bridge.
Another type of bridge is a source routing bridge. This type of bridge
is employed on a token-ring network. A source routing bridge is a
bridge that reads information appended to the packet by the sending
device. This additional information in the packet will state the route
to the destination segment on the network. A source routing bridge
will analyze this information to determine whether or not this
stream of data should or should not be passed along.
Bridges accomplish several things. First, they divide busy networks
into smaller segments. If the network is designed so that most
packets can be delivered without crossing a bridge, traffic on the
individual network segments can be reduced. If the Accounting and
Sales departments are overloading the LAN, for example, you might
divide the network so that Accounting is on one segment and Sales
11.919-x CH06.i
8/28/98 1:04 PM
Page 235
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
on another. Only when Accounting and Sales must exchange packets
does a packet need to cross the bridge between the segments.
Bridges also can extend the physical size of a network. Although the
individual segments still are restricted by the maximum size imposed
by the network design limits, bridges enable network designers to
stretch the distances between segments and extend the overall size of
the network. A general rule of thumb for deciding which side of a
bridge a device should be placed on is that 80% of the device’s traffic should be destined to devices on the same side of the bridge that
the device in question resides on.
Bridges, however, cannot join LANs that are utilizing different network addresses. This is because bridges operate at the Data Link
layer of the OSI model and depends on the physical addresses of
devices and not at the Network layer which relies on logical network
addresses.
235
<$Ibridges;extending size of network><$Ibridges;t
ypes;remote><$Ire
mote bridges;use
with synchronous
modems><$Isynch
ronous
modems;remote
bridges>
Bridges sometimes are also used to link a LAN segment through a
synchronous modem connection to another LAN segment at a
remote location. A so-called remote bridge minimizes modem traffic
by filtering signals that won’t need to cross the modem line (see
Figure 6.8).
A&B Packets
B Packets Only
FIGURE 6.8
A remote bridge acts as a filter for a synchronous modem.
LAN
A
Remote
Bridge
Synchronous
Modem
ROUTING
An internetwork consists of two or more physically connected independent networks that are able to communicate. The networks that
make up an internetwork can be of very different types. For example, an internetwork can include Ethernet and token-ring networks.
11.919-x CH06.i
236
8/28/98 1:04 PM
C h a p te r 6
Page 236
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
Because each network in an internetwork is assigned an address, each
network can be considered logically separate; that is, each network
functions independently of other networks on the internetwork.
Internetwork connectivity devices, such as routers, can use network
address information to assist in the efficient delivery of messages.
Delivering packets according to logical network address information
is called routing. The common feature that unites internetwork connectivity devices (routers and brouters) is that these devices can perform routing. The following list details some common internetwork
connectivity devices:
<$Inetworks;routing;overview><$Ii
nternetworks;connectivity
devices;routers><$
Iinternetworks;con
nectivity
devices;brouters>
á Routers
á Brouters
Each of these devices is discussed in the following sections.
<$Ibridges;versus
routers><$Irouters;ve
rsus
bridges><$Ibridges;li
Node A
mitations>
Node=B
Node=B
Net A
Bridge A
Bridge B
Node=B
Node=B
Node B
FIGURE 6.9
A complex network with bridges.
Routers
Bridges are suitable for relatively simple networks, but bridges have
certain limitations that become more significant in complex network
situations. One limitation of bridges is that packets intended for all
people on a subnet, also known as a broadcast, are received by every
single device on the network. By being able to section off a LAN
segment into different network segments, routers allow you to control and group devices that work together to be on the same network
segment.
Consider the network in Figure 6.9. Both bridges are aware of the
existence of Node B, and both can pick up the packet from Net A
and forward it. At the very least, the same packet can arrive twice at
Node B.
A worse case, however, is that these relatively unintelligent bridges
can start passing packets around in loops, which results in an everincreasing number of packets that circulate on the network and
never reach their destinations. Ultimately, such activity can (and
will) saturate the network.
An algorithm, called the spanning tree algorithm, enables complex
Ethernet networks to use bridges while redundant routes exist. The
algorithm enables the bridges to communicate and construct a logical network without redundant paths. The logical network is reconfigured if one of the paths fails.
8/28/98 1:04 PM
Page 237
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
Another problem is that the bridges cannot analyze the network to
determine the fastest route over which to forward a packet. When
multiple routes exist, this is a desirable capability, particularly in
wide area networks (WANs), where some routes are often considerably slower than others.
Routers organize the large network in terms of logical network segments. Each network segment is assigned an address so that every
packet has both a destination network address and a destination
device address.
Routers are more “intelligent” than bridges. Not only do routers
build tables of network locations, but they also use algorithms to
determine the most efficient path for sending a packet to any given
network. Even if a particular network segment isn’t directly attached
to the router, the router knows the best way to send a packet to a
device on that network. In Figure 6.10, for example, Router A
knows that the most efficient step is to send the packet to Router C,
not Router B.
Notice that Router B presents a redundant path to the path Router
A provides. Routers can cope with this situation because they
exchange routing information to ensure that packet loops don’t
occur. In Figure 6.10, if Router A fails, Router B provides a backup
message path, thus making this network more robust.
One consequence of all the processing a router performs on a packet
is that routers generally are slower than bridges.
You can use routers to divide large, busy LANs into smaller segments, much as you can use bridges. But that’s not the only reason
to select a router. Routers also can connect different network types.
An example of this would be a router that connected a token-ring
segment with the Ethernet segments. On such networks, a router is
the device of choice, as a bridge cannot perform this function.
The protocols used to send data through a router must be specifically designed to support routing functions. IP, IPX, and DDP (the
AppleTalk Network-layer protocol) are routable transport protocols.
NetBEUI is a non-routable transport protocol. Transport protocols
will be discussed in greater detail in Chapter 7.
Because routers can determine route efficiencies, they usually are
employed to connect a LAN to a wide area network (WAN). WANs
frequently are designed with multiple paths, and routers can ensure
that the various paths are used most efficiently.
NOTE
11.919-x CH06.i
237
Internetworks Recall that an internetwork consists of two or more logically separate but physically connected networks. By this definition, any
network segmented with routers is an
internetwork.
<$Ibridges;spanning tree
algorithms><$Ispanning
tree
algorithms><$Irouters;versus bridges><$Ibridges;versus
routers><$Iinternetworks;d
efined><$Irouters;packets;t
ransmission
efficiency><$Ipackets;route
r efficiency versus bridges>
11.919-x CH06.i
238
8/28/98 1:04 PM
C h a p te r 6
Page 238
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
FIGURE 6.10
An internetwork: A series of networks separated by routers.
Net=E, Node=B
Net A
Net B
Router A
<$Iprotocols;routers;support
of><$Irouters;protocols;support
of><$Irouters;situational
uses><$Irouters;types;static><$I
routers;types;dynamic><$Idyna
mic routers><$Istatic routers>
Router B
Router C
Net=E, Node=B
Net C
Net D
Router E
Router D
Net=E, Node=B
Net E
Net E, Node B
The Network layer functions independently of the physical cabling
system and the cabling system protocols—independently, that is, of
the Physical and Data Link layers. This is the reason that routers easily can translate packets between different cabling systems. Bridges,
on the other hand, cannot translate packets in this way because they
function at the Data Link layer, which is closely tied to physical
specifications.
Routers come in two general types:
á Static Routers. These routers do not determine paths. Instead,
you must configure the routing table, specifying potential
routes for packets.
á Dynamic Routers. These routers have the capability to deter-
mine routes (and to find the optimum path among redundant
routes) based on packet information and information obtained
from other routers.
11.919-x CH06.i
8/28/98 1:04 PM
Page 239
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
To determine the best path for a packet, routers employ some form
of routing algorithm. Some common routing algorithms are discussed in the following sections.
239
<$Ibrouters;defined><$I
bridges;brouters><$Irou
ters;brouters>
Brouters
A brouter is a router that also can act as a bridge. A brouter
attempts to deliver packets based on network protocol information,
but if a particular Network layer protocol isn’t supported, the
brouter bridges the packet using device addresses.
GATEWAYS
The term “gateway” originally was used in the Internet protocol
suite to refer to a router. Today, the term “gateway” more commonly
refers to a system functioning at the top levels of the OSI model
that enables communication between dissimilar protocol systems. A
gateway generally is dedicated to a specific conversion, and the exact
functioning of the gateway depends on the protocol translations it
must perform. Gateways commonly function at the OSI Application
layer, but actually can operate at any level of the OSI model.
<$Igateways;OSI Application
layer><$Igateways;conversion
function of><$Igateways;implementation;hardware><$Igatewa
ys;implementation;software>
Gateways connect dissimilar environments by removing the layered
protocol information of incoming packets and replacing it with the
packet information necessary for the dissimilar environment (see
Figure 6.11).
Network A
Network B
Packets in
Network B’s
Format
Packets in Network A’s
Format
FIGURE 6.11
Gateways convert protocol information to dissimilar environments.
GATEWAY
11.919-x CH06.i
240
8/28/98 1:04 PM
C h a p te r 6
Page 240
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
Gateways can be implemented as software, hardware, or a combination of both. An example of a gateway is often seen in email systems. When you send email, say from Microsoft Exchange to someone on the Internet, a gateway is responsible for converting the
Microsoft Exchange message contents and addressing, to one that
is compatible with the SMTP (Internet) message format and
addressing.
DYNAMIC ROUTING APPLIED—
ROUTING ALGORITHMS
<$Inetworks;routing;cost
estimations><$Irouters;cos
t factors;hop
count><$Irouters;cost factors;tic
count><$Irouters;cost factors;relative
expense><$Irouters;cost
factors;static
selection><$Irouters;cost
factors;dynamic selection><$Ihop counts
(routers)><$Itic counts
(routers)><$Idynamic
routers;cost estimation>
Routing refers to the process of forwarding messages through
internetworks of LANs. In some cases, routing information is programmed into the routing devices. However, preprogrammed, or
static, routers cannot adjust to changing network conditions. Most
routing devices, therefore, are dynamic, which means that they have
the capability of discovering routes through the internetwork and
then storing the route information in route tables.
Route tables do not store only path information. They also store
estimates of the time, cost or calculated distance taken to send a
message through a given route. This time estimate is known as the
cost of a particular path. Some of the methods of estimating routing
costs are as follows:
á Hop count. This method describes the number of routers that a
message might cross before it reaches its destination. If all hops
are assumed to take the same amount of time, the optimum
path is the path with the smallest hop count.
á Tic count. This method provides an actual time estimate, where
a tic is a time unit as defined by the routing implementation.
á Relative expense. This method calculates any defined measure of
the cost (including the monetary cost) to use a given link.
After costs are established, routers can select routes, either statically
or dynamically, as follows:
á Static route selection. This selection method uses routes that
have been programmed by the network administrator.
11.919-x CH06.i
8/28/98 1:04 PM
Page 241
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
á Dynamic route selection. Under this selection method, routing
cost information is used to select the most cost effective route
for a given packet. As network conditions change and are
reflected in routing tables, the router can select different paths
to maintain low costs.
Two common methods of discovering routes are distance vector
routing and link-state routing. Both are discussed in the following
sections. It should be mentioned that some networks are too large to
implement dynamic routing, as the routing tables would get too
large. In cases like this, reliance on static routing tables (those that
are manually inputted) is needed. The Internet is a great example of
systems that rely on static routing tables.
<$Inetworks;routing;distance vector><$Inetworks;rout
ing;linkstate><$Idistance
vector routing;inefficiencies><$Idistance
vector routing;advertising availability>
Distance Vector Routing
Distance vector routers advertise their presence to other routers on
the network. Periodically, each router on the network broadcasts the
information in its routing table. Other routers can use this information to update their own router tables.
Figure 6.12 illustrates how the process works. In the figure, Server
S3 learns that Server S2 can reach Server S1 in one hop. Because S3
knows that S2 is one hop away, S3 knows that its cost to reach S1
through S2 is two hops.
Distance vector routing is an effective algorithm, but it can be fairly
inefficient. Because changes must ripple through the network from
S1
S2 IS
ONE HOP
AWAY FROM
ME, SO S1 IS
TWO HOPS
AWAY.
SI IS
ONE HOP
AWAY FROM
ME!
S2
241
S3
FIGURE 6.12
Distance vector routing.
11.919-x CH06.i
242
8/28/98 1:04 PM
C h a p te r 6
Page 242
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
router to router, it might take a while for a change to become known
to all routers on the network. In addition, the frequent broadcasts of
routing information produce high levels of network traffic that can
hurt performance on larger networks.
<$Inetworks;routing;linkstate><$Ilink-state
routing;update process>
Link-State Routing
Link-state routing reduces the network traffic required to update
routing tables. Routers that are newly attached to the network can
request routing information from a nearby router.
After routers have exchanged routing information about the network, routers broadcast messages only when something changes.
These messages contain information about the state of each link the
router maintains with other routers on the network. Because routers
keep each other updated, complete network routing updates are not
needed often.
C A S E S T U DY : B R I D G E S
OR
ROUTERS?
ESSENCE OF THE CASE
• Should this company use bridges or
routers?
you are to address these issues where applicable when making your presentation to the planning team. Your job is to present a brief summary of the benefits and costs of using bridges or
routers.
• Explain the benefits and costs of each
device.
A N A LY S I S
The essence of the case is as follows:
• Do not focus on the details such as
topology or cable types but raise these
issues where applicable.
SCENARIO
You are involved in the planning stages of a network that is about to be placed within a company. There is concern over whether to utilize
bridges or routers within the network. Actual
cabling types, topologies, and protocols are not
the primary concern at this planning event, but
If you were in this situation you would need to
explain the function of a bridge and a router to
the company. You should also make sure that
this company is aware that bridges and routers
compliment each other; they are not necessarily
replacements for each other.
Bridges have the following features. They are
transport protocol independent (see Chapter 7
for more information on transport protocols).
Bridges also are not management intensive, as
they usually in most cases do not require any
configuration information. They will learn the
11.919-x CH06.i
8/28/98 1:04 PM
Page 243
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
C A S E S T U DY : B R I D G E S
OR
243
ROUTERS?
network segment, once they are attached.
Bridges are often cheaper to purchase than a
router. The main purpose of using bridges in an
Ethernet network is to reduce collisions, where
as in a token-ring network, they are used to
bridge rings together. A possible replacement for
a bridge could be a switch, also known as a
switched hub. However, switched hubs tend to
cost more than a bridge.
The problem with a bridge is that a bridged network is reliant on the transmission medium properties as to the number of nodes that can be on
the network. Also, bridges are ideally supposed
to follow the 80/20 industry rule of thumb. That
is 80% of a node’s traffic should be on its side of
the bridge. Only 20% of a node’s traffic should go
to the other side of the bridge. A bridge has no
method of locating devices either. That is,
because it relies only on MAC addresses, and
not on any logical grouping address (network
address), bridges are not ideal for large networks.
Routers require more manual intervention than
bridges. They often cost more, and they are
dependent upon transport protocols in order to
function. The ideal thing about routers is that
they can connect different LANs together to form
an internetwork. Thus one can have a network of
several different transmission media and topologies. One can also connect more devices together, utilizing different LANs, than possible when
only having one network segment.
In large networks, routers are most often used
in order to logically group devices, or because
different transmission media are being used to
connect different segments of a network. Bridges
are often placed on the individual network
segments of a network in order to reduce collisions on that particular network segment.
One issue that would involve the use of routers
would be if the router was to be static or dynamic. A dynamic router would need to use some
form of dynamic routing protocol. Once a selected protocol was to be established, a particular
router from a vendor could be selected.
In short, characteristics of bridges are:
• Bridges are cheaper than routers.
• Bridges usually do not require any manual
intervention.
• Bridges do not extend the number of
devices that can go on a transmission
medium.
• Bridges have no method of locating devices
on the cable segment.
• Learning bridges are used on Ethernet networks, whereas Source Routing bridges are
used on token-ring networks.
Issues concerning routers are:
• Routers cost more than bridges.
• Routers often will require manual intervention to configure them.
• Routers will connect different LAN segments together, allowing a network to grow
beyond the limitations of one transmission
medium.
• Routers can use logical addresses to locate
devices on the network.
• Routers can be used to connect Ethernet to
token-ring networks.
11.919-x CH06.i
244
8/28/98 1:04 PM
C h a p te r 6
Page 244
C O NNE C TI V I TY D E VICES AND
<$IChapter 6;key terms><$Ikey terms;Chapter
TRANSFER
MECHANISMS
6>
CHAPTER SUMMARY
KEY TERMS
• Internetwork
• Modem
• PSTN
• Asynchronous Modems
• Synchronous Modems
• Concentrator
• Active Hub
• Intelligent Hub
• Transparent Bridge
• Learning Bridge
• Spanning Tree Algorithm
• Static Router
• Dynamic Router
• Hop count
• Tic count
• Relative expense
• Distance Vector Routing
• Link State Routing
• Repeaters
• Bridges
• Routers
• Brouter
• Gateways
This chapter examined some of the connectivity devices that network engineers use to expand, optimize, and interconnect networks.
These devices have some similarities, but each is designed for a specific task, as described in the following list:
á Repeaters. Repeaters regenerate a signal and are used to expand
LANs beyond cabling limits.
á Bridges. Bridges know the side of the bridge on which a node
is located. A bridge passes only packets addressed to computers
across the bridge, so a bridge can thus filter traffic, reducing
the load on the transmission medium.
á Routers. Routers forward packets based on a logical (as
opposed to a physical) address. Some routers can determine
the best path for a packet based on routing algorithms.
á Gateways. Gateways function under a process similar to routers
except that gateways can connect dissimilar network environments. A gateway replaces the necessary protocol layers of a
packet so that the packet can circulate in the destination environment.
You should be familiar with the features of these connectivity
devices and with their relative advantages and disadvantages for the
Networking Essentials exam.
This chapter also analyzed several forms of transmission used by
devices such as network cards and modems, both asynchronous and
synchronous, and concluded with different types of routing algorithms used by routers, those being Distance Vector and Link State
routing.
There are two major points to keep in mind about the material in
this chapter. The first is that the devices discussed, particularly
repeaters, bridges and routers, are not competing technologies but
complementary technologies. The second is that, although the exam
objective does not specifically identify transmission types and routing algorithms, it is important to be aware of them as a network
professional. These topics were included in this chapter because it is
the devices in this chapter that use these routing algorithms and
transmission methods.
11.919-x CH06.i
8/28/98 1:04 PM
Page 245
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
245
A P P LY Y O U R L E A R N I N G
Exercises
6.2
6.1
Objective: Learn how to configure parameters on an
asynchronous modem within Windows 95.
Equipment Costs
Objective: Learn the costs of different hardware used
on a network.
Configuring an Asynchronous Device
Estimated Time: 10 minutes
1. Select the Start button, move to the Settings
option and then select the Control Panel option.
Estimated Time: 10 minutes
The purpose of this exercise is to call up a vendor and
find out the prices of different components used on the
network. Possible companies that you may ask for, that
produce these devices are Cisco, Bay Networks, Intel
and 3Com, to name but a few.
2. Double click on the modem icon.
The main purpose is to find out the difference between
the prices of a Repeater, Hub, and Router. If the company you call does not have the exact device as listed,
substitute with one that is close.
4. From the Install New Modem dialog box, check
Don’t detect my modem; I will select it from a
list option and then click on the Next button.
3. If you have a modem already installed, proceed to
step 8. If you do not have a modem installed, a
modem installer wizard opens up.
5. From the Install New Modem dialog box, select
the “(Standard Modem Types)” option under the
Manufacturers list, and also select the Standard
28800bps Modem option under Types list. Then
click on the Next button. (As we will not actually
use the modem, it does not matter if you actually
have a physical modem attached to your PC.)
When asking about the Routers, ask if these are passive
or active routers. If the quotes you receive are for active
routers, ask which routing algorithms the router supports.
Use the following table:
Item
Description
Price
Repeater
- 10BASE-T Cable
$_________
- 10BASE-2 Cable
$_________
- 20 port passive (Ethernet)
$_________
- 20 port active (Ethernet)
$_________
- 20 port switched hub (Ethernet)
$_________
MAU
- 20 port MAU for token-ring
$_________
Router
- 3 connections, all for Ethernet
$_________
- 1 ISDN connection and one
token-ring connection
$_________
HUB
<$IChapter
6;exercises><$Iexercises;Chap
ter 6>
11.919-x CH06.i
246
8/28/98 1:04 PM
C h a p te r 6
Page 246
C O NNE C TI V I TY D E VICES
<$IChapter 6;exercises><$Iexercises;Chapter
6;review questions><$Ireview quesAND 6><$IChapter
TRANSFER MECHANISMS
tions;Chapter 6><$IChapter 6;exam
A P P LY Y O U R L E A R N I N G
6. In the Install New Modem dialog box, select
Communications Port (Com3) under the Select
the port to use with this modem list. Then click
on the Next button.
7. Click on the Finish button.
8. In the Modem Properties dialog box select the
Standard 28800bps Modem or your previously
installed modem. Then click on the Properties
button.
9. Select the Connection tab.
Examine the Connection preferences options. It is here
that you can adjust for how many stop bits are to be
used as well as the number of data bits and the parity
type. All of these are options that control the amount
of overhead used in an asynchronous communication
method.
Exam Questions
1. Your LAN includes computers in two rooms at
different ends of the company office. The cables
connecting the rooms exceed the maximum
cabling distance for the transmission medium,
and the network is experiencing problems due to
signal loss in the long cables. What would be the
cheapest and simplest solution?
A. Router
B. Repeater
C. Bridge
D. Brouter
2. Your Ethernet LAN is experiencing performance
problems due to heavy traffic. What would be a
simple solution?
A. Gateway
Review Questions
1. When discussing Repeaters, Bridges, and Routers,
which device relies on logical network addresses,
physical hardware addresses, or does not use
addresses at all?
2. Describe the difference between synchronous and
asynchronous transmissions.
3. Describe the difference between static and
dynamic routing.
4. What is the difference between distance vector
routing and link-state routing?
B. Repeater
C. Bridge
D. Router
3. What routing algorithm enables bridges to operate on a network with redundant routes?
A. Distance vector
B. Link-state
C. Spanning tree
D. Learning tree
11.919-x CH06.i
8/28/98 1:04 PM
Page 247
<$IChapter 6;exam questions><$Iexam quesCh a p t e r 6 CONNECTIVITY
tions;Chapter 6>
DEVICES AND TRANSFER MECHANISMS
247
A P P LY Y O U R L E A R N I N G
4. You need to connect a token-ring LAN with an
Ethernet LAN. To do so, you will need what type
of device?
A. Bridge
B. Gateway
C. Repeater
D. Hub
5. You need to connect a token-ring and an
Ethernet LAN segment. To do so, you will need
what type of device?
8. Which networks can use MAUs?
A. Ethernet
B. ARCnet
C. Token-ring
D. All the above
9. Which two of the following features can add
intelligence to a hub?
A. Signal regeneration
B. Network-management protocols
A. Repeater
C. Multiport repeaters
B. Bridge
D. Switching circuitry
C. Remote bridge
D. Router
6. What device uses an address table to determine
where to send a packet?
A. Bridge
10. Which two statements are true of repeaters?
A. Repeaters filter network traffic.
B. Repeaters extend network distances.
C. Repeaters regenerate signals.
D. Repeaters operate at the OSI Data Link layer.
B. Router
C. Both A and B
D. None of the above
7. Which three of the following are advantages of
active hubs over passive hubs?
A. They can regenerate network signals.
B. LAN ranges can be extended.
C. They are less expensive.
D. They function as repeaters.
11. Which three statements are true of bridges?
A. Bridges amplify and regenerate signals.
B. Bridges can connect logically separate networks.
C. Bridges use device address tables to route
messages.
D. Bridges divide networks into smaller segments.
11.919-x CH06.i
248
8/28/98 1:04 PM
C h a p te r 6
Page 248
C O NNE C TI V I TY D E VICES AND
<$IChapter 6;exam questions><$Iexam questions;Chapter
6>
TRANSFER
MECHANISMS
A P P LY Y O U R L E A R N I N G
12. Which of the following connectivity devices is
the least expensive?
A. Repeater
B. Bridge
C. Router
16. On what level of the OSI does a hub reside?
A. Physical
B. Network
C. Application
D. Transport
D. Gateway
17. On what level of the OSI does a bridge reside?
13. Which of the following connectivity devices uses
logical addresses?
A. Repeater
B. Bridge
C. Router
D. None of the above
14. Which of the following connectivity devices connects dissimilar networking protocol environments?
A. Repeater
A. Data Link
B. Physical
C. Network
D. Presentation
18. At what level does a Router live?
A. Network
B. Physical
C. Application
D. Data Link
B. Bridge
C. Router
D. Gateway
19. At what level of the OSI does a Gateway live?
A. Physical and Data Link
B. Physical
15. What type of router requires a human-configured
routing table?
A. Explicit router
B. Satic router
C. Simple router
D. Bridge
C. It can operate at multiple levels throughout
the OSI model.
D. Network and Data Link
11.919-x CH06.i
8/28/98 1:04 PM
Page 249
<$IChapter 6;answers><$Ianswers;Chapter 6>
Ch a p t e r 6
CONNECTIVITY DEVICES AND TRANSFER MECHANISMS
249
A P P LY Y O U R L E A R N I N G
Answers to Review Questions
1. Repeaters operate at the Physical layer of the OSI
model. Repeaters do not rely on any type of
address at all, as their function is to simply regenerate a signal, regardless of where the signal is
destined to go.
Bridges operate at the Data Link layer of the OSI
model. Bridges analyze the hardware addresses, or
MAC addresses, of the source and destination
devices on each packet. Based upon these MAC
addresses, a packet will be allowed through a
bridge if that device is on the other side, or not
through a bridge if the device is not on the other
side of the bridge.
Routers utilize logical network addresses to define
separate network segments. These logical network
addresses are usually assigned by the network
administrator, and will follow specific naming
conventions, depending upon the protocol being
used.
See the sections titled “Repeaters,” “Bridges,” and
“Routing.”
2. Asynchronous transmission utilizes a start and a
stop bit when sending data, as there is no clocking being done. Up to 30% of the signals being
transmitted can be the overhead of the start bit,
stop bit, and the error correcting mechanism.
Modems use asynchronous transmissions.
Synchronous transmissions, allow for the removal
of the overhead associated with asynchronous
transmissions. It can accomplish this by encoding
clocking signals in the data, or by using a separate communication channel to send the clocking
signal. Network cards utilize synchronous transmissions.
See the sections titled “Asynchronous
Transmission” and “Synchronous Transmission.”
3. Static routing requires manually entering routing
tables into a router, so that the router knows
which paths house different networks.
Dynamic routers share routing information with
each other automatically. Routers do this by using
either a link-state or distance vector routing protocol.
See the section titled “Routing.”
4. Distance vector routing uses broadcasts at periodic intervals to announce the presence of the
routes to other routers. Distance vector routing is
effective, but can produce significant amounts of
traffic on a large network.
In link state routing, newly attached routers can
request information, as soon as something on the
network changes. It is an ideal routing protocol
to use on a large network, but usually requires
more manual configuration than distance vector
routing.
See the section titled “Dynamic Routing
Applied—Routing Algorythms.”
Answers to Exam Questions
1. B. Routers, brouters, and bridges are designed to
isolate traffic, not regenerate signals to exceed
cable distance recommendations. See the section
titled “Repeaters.”
2. C. Bridges and routers are solutions to isolate
traffic. A bridge is a more simple solution than a
router as bridges often do not need to be configured. See the section titled “Bridges.”
11.919-x CH06.i
8/28/98 1:04 PM
Page 250
<$IChapter 6;answers><$Ianswers;Chapter 6>
250
C h a p te r 6
C O NNE C TI V I TY D E VICES AND TRANSFER MECHANISMS
A P P LY Y O U R L E A R N I N G
3. C. A and B are router algorithms. D just sounds
good. See the sections titled “Bridges” and
“Routers.”
11. A, C, D. B is incorrect as this is the function of a
router. See the section titled “Bridges” and the
one titled “Routers.”
4. B. Gateways are what are used to connect dissimilar systems. See the section titled “Gateways.”
12. A. Repeaters tend to be the cheapest of all four
options. See the section titled “Repeaters.”
5. D. See the answer above.
13. C. Only routers use logical addresses. See the sections titled “Routing” and “Addressing.”
6. C. Both bridges and routers consult an address
table when determining where a packet is to be
sent. For more information see the sections titled
“Bridges” and “Routers.”
7. A, B, D. Active hubs are more expensive than
passive hubs. See the section titled “Hubs.”
8. C. MAUs are used in token-ring networks. See
the section titled “Hubs.”
9. B, D. A and C are functions of a repeater, not an
Intelligent hub. See the section titled “Hubs.”
10. B, C. D is incorrect as a repeater operates at the
Physical level of the OSI model. A is incorrect as
repeaters do not filter traffic. See the section titled
“Repeaters.”
14. D. Gateways are responsible for connecting dissimilar networking environments. See the section
titled “Gateways.”
15. B. There are two types of routers, static and
dynamic. Static routers require human intervention. See the section titled “Routers.”
16. A. Hubs operate at the Physical layer of the OSI
model. See the section titled “Hubs.”
17. A. Bridges operate at the Data Link layer of the
OSI model. See the section titled “Bridges.”
18. A. Routers operate a the the Network layer of the
OSI model. See the section titled “Routing.”
19. C. Gateways operate at many different layers of
the OSI model. See the section titled “Gateways.”
Suggested Readings and Resources
1. Derfler, Frank J., Jr. Using Networks. Que,
1998.
• Chapter 10: Lan Portals
2. Ford, M. Internetworking Technologies
Handbook. Macmillan Technical Publishing,
1997.
12.919-x CH07.i
8/28/98 1:10 PM
Page 251
OBJECTIVES
Chapter 7 targets the following objectives in the
Planning and Implementation sections of the
Networking Essentials exam:
Select the appropriate network and transport protocols for various token-ring and Ethernet networks. Protocols include DLC, AppleTalk, IPX,
TCP/IP, NFS, and SMB.
. When devices communicate over the network, they
must utilize some form of transport protocol or set
of rules to move data from one device to another.
This exam objective reflects the need for you to
know the transport protocols that are used most
often with Windows 95 and Windows NT.
Implement a NetBIOS naming scheme for all computers on a given network.
. Microsoft networking components rely on the
capability to reference other machines on the network using NetBIOS names. This exam objective
makes it clear that you must have the ability to deal
with NetBIOS naming rules for computers.
C H A P T E R
7
Transport Protocols
12.919-x CH07.i
8/28/98 1:10 PM
Page 252
OUTLINE
S T U DY S T R AT E G I E S
Packets and Protocols
254
Protocols and Protocol Layers
256
TCP/IP—Internet Protocols
General TCP/IP Transport Protocols
TCP/IP Services
TCP/IP Routing Protocols
257
259
263
266
NetWare IPX/SPX
266
General IPX/SPX Transport Protocols 267
IPX/SPX Services
272
IPX/SPX Routing
272
NetBEUI
273
AppleTalk
273
Data Link Control (DLC)
275
Relating Protocol Stacks Together
Server Messaging Blocks (SMB)
NetBIOS Names
275
276
277
NetBIOS Background
277
Assigning NetBIOS Names
277
Finding Resources on Microsoft
Networks
280
Chapter Summary
284
. This chapter presents two important exam topics. The first topic is concerned with different
protocols, or sets of rules, used in networking.
The second objective focuses on NetBIOS
names and naming conventions used by
Microsoft’s networked operating systems.
. To give a full account of the first exam topic,
many different protocols are presented. In
preparing for the exam, pay particular attention
to the features and functions of each of the six
protocols listed. But also be aware of what the
other protocols accomplish to ensure you do
not confuse one protocol with another.
. With the second exam topic on NetBIOS names,
be aware of what characters are not allowed,
and what general naming conventions are followed.
12.919-x CH07.i
8/28/98 1:10 PM
Page 253
C hapter 7
INTRODUCTION
In Chapter 2, “Networking Standards,” you learned that the design
of network protocols is usually done in pieces, with each piece
solving a small part of the overall problem. By convention, these
protocols are regarded as layers of an overall set of protocols, called a
protocol suite or a protocol stack. A protocol stack often covers the
entire OSI reference model.
As Chapter 2 describes, the OSI reference model is a standard describing the activities at each level of a protocol stack. The OSI reference
model is useful as a conceptual tool for understanding protocol layering. Chapters 3 through 6 have built upon the OSI reference
model from the bottom up. This chapter discusses in detail the components that operate at the Network layer through the Application
layer.
This chapter examines a variety of actual transport protocols and
protocol suites, such as TCP/IP and IPX/SPX. Although some protocol stacks have been designed in strict conformance with the OSI
reference model, full OSI compliance is not usually the norm. Many
of these protocol stacks have their origins in the days before the OSI
model, and thus can be matched only loosely to the seven-layer OSI
model. The main use of the OSI reference model is as a conceptual
framework for understanding network communication and comparing various types of protocols.
This chapter begins by reviewing and placing into context the information learned from the previous chapters. The analysis begins with
an examination of packets and protocols, as well as protocols and
their reference back to the OSI model. From that point, the transport protocols of TCP/IP, IPX/SPX, NetBEUI, AppleTalk, and DLC
are examined. When analyzing these transport protocols, issues such
as addressing, routing mechanisms, and services are addressed. The
chapter continues by examining NetBIOS naming schemes that are
used in Microsoft networks. From there, networking as a whole is
applied to the Windows NT model.
TRANSPORT PROTOCOL S
253
12.919-x CH07.i
254
8/28/98 1:10 PM
C h a p te r 7
Page 254
TR A NS P O R T P ROTOCOLS
PACKETS
AND
PROTOCOLS
NOTE
Before investigating protocols and protocol stacks, take a moment to
quickly review some of the protocol-related issues discussed in previous chapters.
Standards and Protocols The NDIS
and ODI standards greatly simplify the
task of finding common protocols.
NDIS and ODI (described in Chapter
2) enable several transport protocols
to operate simultaneously through the
same network adapter card.
The purpose of a network is to exchange information among computers, and protocols are the rules by which computers communicate. Computers, like humans, can adopt any number of systems for
passing messages, as long as the sending and receiving computers are
using the same (or compatible) rules. Computers, therefore, must
agree on common protocols before they can communicate. Failing to
do so would create a bewildering situation similar to what you’d face
if you read a book in Russian to a listener who speaks only
Cherokee.
You can classify the many tasks that network protocols must oversee
into a few basic categories. Think of these categories chronologically,
as a series of steps (each step including a collection of related tasks)
that must take place before the data can reach the transmission
medium. These steps are the layers of a protocol stack, as described
in Chapter 2. In one sense, the term layer is more than metaphorical.
Each layer of the stack (the Application layer, the Presentation layer,
and so on) adds a layer of information to the packet; the corresponding layer of the receiving computer needs to process the incoming
packet.
The purpose of the layering structure is to enable vendors to adapt
to specific hardware and software components without having to
recreate the entire protocol stack.
Protocols describe the way in which network data is encapsulated in
packets on the source end, sent via the network to a destination, and
then reconstructed at the destination into the appropriate file,
instruction, or request. Breaking network data into packet-sized
chunks provides smoother throughput because the small packets
don’t tie up the transmission medium as a larger unit of data might.
Also, packets simplify the task of error detection and correction.
Each file is checked separately for errors, and if an error is discovered, only that packet (rather than a whole file) must be retransmitted.
The exact composition of a network packet depends on the protocols you’re using. In general, network packets contain the following:
12.919-x CH07.i
8/28/98 1:10 PM
Page 255
C hapter 7
TRANSPORT PROTOCOL S
255
á Header. The header signifies the start of the packet and con-
tains a bundle of important parameters, such as the source and
destination address and time/synchronization information.
á Data. This portion of the packet contains the original data
being transmitted.
á Trailer. The trailer marks the end of the packet and typically
contains error-checking (Cyclical Redundancy Check, or
CRC) information.
As the data passes down through the protocol layers, each layer
performs its prescribed function, such as interfacing with an application, converting the data format, or adding addressing and errorchecking parameters. (Chapter 2 examines the functions of the OSI
protocol layers.) As you learn in this chapter, actual working protocol stacks don’t always comply exactly with the OSI model—some,
in fact, predate the OSI model—but the concepts and terminology
of the OSI model are nevertheless useful for describing protocol
functions.
When the packet reaches the transmission medium, the network
adapter cards of other computers on the network segment examine
the packet, checking the packet’s destination address. If the destination address matches the PC’s address, the network adapter interrupts the processor, and the protocol layers of the destination PC
process the incoming packet (see Figure 7.1).
COMPUTER
A
PROCESSOR
Yes
For
A?
ALL
PACKETS
No
IGNORE
FIGURE 7.1
The network adapter card checks whether the
destination address matches the PC’s address.
12.919-x CH07.i
256
8/28/98 1:10 PM
C h a p te r 7
Page 256
TR A NS P O R T P ROTOCOLS
PROTOCOLS
AND
PROTOCOL LAYERS
Many of the addressing, error-checking, retransmission, and
acknowledgment services most commonly associated with networking take place at the Network and Transport OSI layers. (Refer to
Chapter 2.) Protocol suites are often referred to by the suite’s main
Transport and Network protocols. In TCP/IP, for instance, TCP is a
Transport layer protocol and IP is a Network layer protocol. (Note,
however, that TCP/IP predates OSI and diverges from OSI in a
number of ways.) IPX/SPX is another protocol suite known by its
Transport and Network layer protocols, but the order of the protocols is backward from the way the protocols are listed in TCP/IP.
IPX is the Network and Transport layer protocol; SPX is the
Transport layer protocol.
The lower Data Link and Physical layers of the OSI model provide a
hardware-specific foundation, addressing items such as the network
adapter driver, the media access method, and the transmission medium. Transport and Network layer protocols such as TCP/IP and
IPX/SPX rest on that Physical and Data Link layer foundation, and,
with the help of the NDIS and ODI standards, multiple protocol
stacks can operate simultaneously through a single network adapter.
(Refer to the discussion of NDIS and ODI in Chapter 2.)
Upper-level protocols, those from the Network layer and higher,
allow for the connection of services and the services themselves. This
can imply routing programs, addressing schemes, and File and Print
services.
This chapter describes the common protocol suites and many of the
important protocols associated with them. In addition to TCP/IP
and IPX/SPX, some of the common Transport and Network layer
protocols are the following:
á NWLink. Microsoft’s version of the IPX/SPX protocol essen-
tially spans the Transport and Network layers.
á NetBEUI. Designed for Microsoft networks, NetBEUI
includes functions at the Network and Transport layers.
NetBEUI isn’t routable and therefore doesn’t make full use of
Network layer capabilities.
á AppleTalk Transaction Protocol (ATP) and Name Binding
Protocol (NBP). ATP and NBP are AppleTalk Transport layer
protocols.
12.919-x CH07.i
8/28/98 1:10 PM
Page 257
C hapter 7
á Data Link Control (DLC). This is used to connect to IBM
Mainframes and Hewlett-Packard JetDirect printers.
TCP/IP—Internet Protocols
Select the appropriate network and transport protocols for various
token-ring and ethernet networks. Protocols include the following:
DLC, AppleTalk, IPX, TCP/IP, NFS, and SMB.
The TCP/IP protocol suite (also commonly called the Internet protocol suite) was originally developed by the United States Department of Defense (DoD) to provide robust service on large internetworks that incorporate a variety of computer types. Part of the main
purpose of this protocol was for it to be hardware-independent. In
some literature, the TCP/IP protocol suite is referred to as the DoD
model. In recent years, the Internet protocols constitute the most
popular network protocols currently in use.
One reason for the popularity of TCP/IP is that no one vendor
owns it, unlike the IPX/SPX, DNA, SNA, or AppleTalk protocol
suites, all of which are controlled by specific companies. TCP/IP
evolved in response to input from a wide variety of industry sources.
Consequently, TCP/IP is the most open of the protocol suites and is
supported by the widest variety of vendors. Virtually every brand of
computing equipment now supports TCP/IP. This has lead to some
problems, though. Because TCP/IP is an open standard, sometimes
one vendor’s implementation of TCP/IP does not work with another’s implementation.
Much of the popularity of the TCP/IP protocols comes from their
early availability on UNIX. The protocols were built into the
Berkeley Standard Distribution (BSD) UNIX implementation. Since
then, TCP/IP has achieved universal acceptance in the UNIX community and is a standard feature on all versions of UNIX.
Figure 7.2 illustrates the relationship of the protocols in the Internet
suite to the layers of the OSI reference model. Notice that the suite
doesn’t include protocols for the Data Link or Physical layers.
TCP/IP was designed to be hardware-independent and thus is able
to work over established standards such as ethernet, token-ring, and
ARCnet, to name but a few lower OSI layer standards. Over time,
TCP/IP has been interfaced to the majority of Data Link and
Physical layer technologies.
TRANSPORT PROTOCOL S
257
12.919-x CH07.i
258
8/28/98 1:10 PM
C h a p te r 7
Page 258
TR A NS P O R T P ROTOCOLS
Application
Presentation
FTP
TELNET
SMTP
NFS
Transport
Network
RIP
NLSP
Session
TCP
UDP
ICMP
IP
DNS
ARP
Data Link
Physical
FIGURE 7.2
TCP/IP or the “Internet Protocol Suite.”
The Internet protocols do not map cleanly to the OSI reference
model. The DoD model was, after all, developed long before the
OSI model was defined. The model for the Internet protocol suite
has four layers (refer to Figure 7.2). From this model, you can see
the approximate relationships of the layers. The DoD model’s layers
function as follows (see Figure 7.3).
á The Network Access layer corresponds to the bottom two layers
of the OSI model. This correspondence enables the DoD protocols to coexist with existing Data Link and Physical layer
standards.
á The Internet layer corresponds roughly to the OSI Network
layer. Protocols at this layer move data between devices on networks.
á The Host-to-Host layer can be compared to the OSI Transport
layer. Host-to-Host protocols enable peer communication
between hosts on the internetwork. (At the time these protocols were designed, personal computers and workstations didn’t
exist, and all network computers were host computers. As a
result, devices on TCP/IP networks are typically referred to as
hosts. The concept of a client/server relationship didn’t exist,
and all communicating hosts were assumed to be peers.)
á The Process/Application layer embraces functions of the OSI
Session, Presentation, and Application layers. Protocols at this
layer provide network services.
One huge advantage of using TCP/IP is that TCP/IP is required for
communication over the Internet; thus the Internet can be used as a
communication backbone. One disadvantage is that the size of the
protocol stack makes TCP/IP difficult to implement on some older
machines. (Present-day PC models should have no problem running
TCP/IP.) TCP/IP has traditionally been considered slower than other
12.919-x CH07.i
8/28/98 1:10 PM
Page 259
C hapter 7
Process/Application
Session
Transport
Host to Host
Network
Internet
Data Link
259
FIGURE 7.3
Application
Presentation
TRANSPORT PROTOCOL S
Physical
Network
Access
OSI
DoD
protocol stacks, because data must be analyzed up to the Network
layer of the OSI model to be evaluated. But again, the power of the
newer machines overcomes much of this difficulty.
A large number of protocols are associated with TCP/IP. These different protocols are grouped into the following unofficial categories:
á General TCP/IP Transport Protocols
á TCP/IP Services
á TCP/IP Routing
Several of these are discussed briefly in the following sections.
TCP/IP is a complex topic, the scope of which runs beyond this
book and the Networking Essentials exam. Microsoft has a full certification exam on their version of TCP/IP. The “Suggested Readings
and Resources” section at the end of this chapter suggests several
good sources on this topic and TCP/IP in general.
General TCP/IP Transport Protocols
This subsection covers general protocols dealing with the addressing
and transportation of packets across the LAN using TCP/IP. All services and routing issues that fall into the TCP/IP protocol stack use
one or more of these Network or Transport layer protocols.
Addressing in TCP/IP
One of the first aspects of transport protocols that needs to be discussed is how the protocols address entities on the network. As discussed several times previously in this book, there are two main
forms of addresses: a node address and a logical network address. A
node address is the address of the entity or device on the network,
A comparison of the TCP/IP layers to the OSI
model.
12.919-x CH07.i
260
8/28/98 1:10 PM
C h a p te r 7
Page 260
TR A NS P O R T P ROTOCOLS
whereas the logical network address is the segment on the network
to which the node is attached.
TCP/IP uses a unique numbering scheme that encapsulates the network and node address into a set of numbers. This number is what
is known as an IP address. All devices on a network that runs the
TCP/IP protocol suite need a unique IP address.
An IP address is a set of four numbers, or octets, that can range in
value between 0 and 255. Each octet is separated by a period. Some
examples are shown here:
á 34.120.66.79
á 200.200.20.2
á 2.5.67.123
á 107.219.2.34
These addresses are actually broken down into three distinct classes.
These are known as class A, class B, and class C addresses.
Class A IP addresses contain a number between 1 and 127 before
the first dot. Some examples are 3.3.6.8, 102.100.77.8, and
23.23.45.67. In a class A address, this first octet represents the network address, and the last three octets represent the node or host
number. Hence an IP address of 69.23.104.200 would represent
host number 23.104.200 on network 69.
Class B and C addresses follow a similar principal to that exemplified in the class A addresses. In the case of a class B address, the first
octet can range in value from 128 to 191, but it is the first two
octets that make up the network address, and the last two octets that
make up the host ID. In the case of a class C address, the first octet
can range in value from 192 to 223, and the first three octets make
up the host ID.
There are class D and E addresses as well. For these addresses, the
first octet is a number greater than 223. These addresses are not currently available to be used and are reserved for other purposes.
In summary, the differences in the classes of the IP addresses reside
in which numbers are to be used in the first octet, which octets represent the Network ID, and which numbers represent the host ID.
Table 7.1 shows three examples, one from each class of address.
12.919-x CH07.i
8/28/98 1:10 PM
Page 261
C hapter 7
TABLE 7.1
CLASSES
AND
ADDRESSES
Class
IP Address
Network ID
Host ID
Class A
102.44.7.100
102.0.0.0
X.44.7.10
Class B
131.107.4.6
131.107.0.0
X.X.4.6
Class C
200.9.88.250
200.9.88.0
X.X.X.250
The topic of TCP/IP addressing goes well beyond the scope of the
information covered in this book. As noted above, it is covered in
more detail in books and courses that focus on TCP/IP.
One last thing to be aware of when discussing IP addresses is the
fact that every device on the network—that is, every computer,
printer, router, or any other device that can be specifically
attached—needs a unique IP address. In other words, everything
needs an IP address and no two IP addresses can be the same in a
given network. If they were, you would end up with two devices on
the network that have the same network and host ID.
Internet Protocol (IP)
The Internet Protocol (IP) is a connectionless protocol that provides
datagram service, and IP packets are most commonly referred to as
IP datagrams. IP is a packet-switching protocol that performs the
addressing and route selection. An IP header is appended to packets,
which are transmitted as frames by lower-level protocols. IP routes
packets through internetworks by utilizing routing tables that are
referenced at each hop. Routing determinations are made by consulting logical and physical network device information, as provided
by the Address Resolution Protocol (ARP).
IP performs packet disassembly and reassembly as required by packet size limitations defined for the Data Link and Physical layers
being implemented. IP also performs error checking on the header
data using a checksum, although data from upper layers is not errorchecked.
Transmission Control Protocol (TCP)
The Transmission Control Protocol (TCP) is an internetwork
connection-oriented protocol that corresponds to the OSI Transport
layer. TCP provides full-duplex, end-to-end connections. When the
TRANSPORT PROTOCOL S
261
12.919-x CH07.i
262
8/28/98 1:10 PM
C h a p te r 7
Page 262
TR A NS P O R T P ROTOCOLS
overhead of end-to-end communication acknowledgment isn’t
required, the User Datagram Protocol (UDP) can be substituted for
TCP at the Transport (host-to-host) level. TCP and UDP operate at
the same layer.
TCP corresponds to SPX in the NetWare environment (see the
“NetWare IPX/SPX” section). TCP maintains a logical connection
between the sending and receiving computer systems. In this way,
the integrity of the transmission is maintained. TCP detects any
problems in the transmission quickly and takes action to correct
them. The trade-off is that TCP isn’t as fast as UDP, due to the
number of acknowledgments received by the sending host.
TCP also provides and assumes message fragmentation and reassembly and can accept messages of any length from upper-layer protocols. TCP fragments message streams into segments that can be
handled by IP. This process enables the application being used to not
break up the data into smaller blocks. IP still can perform fragmentation for UDP packets and further fragmentation for TCP packets.
When used with IP, TCP adds connection-oriented service and performs segment synchronization, adding sequence numbers at the
byte level.
In addition to message fragmentation, TCP can multiplex conversations with upper-layer protocols and can improve use of network
bandwidth by combining multiple messages into the same segment.
Each virtual-circuit connection is assigned a connection identifier
called a port, which identifies the datagrams associated with that
connection.
User Datagram Protocol (UDP)
The User Datagram Protocol (UDP) is a connectionless Transport
(host-to-host) layer protocol. UDP does not provide message
acknowledgments; rather, it simply transports datagrams.
Like TCP, UDP utilizes port addresses to deliver datagrams. These
port addresses, however, aren’t associated with virtual circuits and
merely identify local host processes. UDP is preferred over TCP
when high performance or low network overhead is more critical
than reliable delivery. Because UDP doesn’t need to establish,
maintain, and close connections, or control data flow, it generally
outperforms TCP. The downfall in UDP is that it does not perform
as reliably as TCP when transmitting data; thus, UDP is often used
when transmitting smaller amounts of data.
12.919-x CH07.i
8/28/98 1:10 PM
Page 263
C hapter 7
UDP is the Transport layer protocol used with the Simple Network
Management Protocol (SNMP), the standard network management
protocol used with TCP/IP networks. UDP enables SNMP to provide network management with a minimum of network overhead.
Address Resolution Protocol (ARP)
Three types of address information are used on TCP/IP internetworks:
á Physical addresses. Used by the Data Link and Physical layers.
á IP addresses. Provide logical network and host IDs. IP
addresses consist of four numbers typically expressed in
dotted-decimal form.
á Logical node names. Identify specific hosts with alphanumeric
identifiers, which are easier for users to recall than the numeric
IP addresses. An example of a logical node name is
MYHOST.COM.
Given an IP address, the Address Resolution Protocol (ARP) can
determine the physical address used by the device containing the IP
address. ARP maintains tables of address resolution data and can
broadcast packets to discover addresses on the network segment or
use previously cached entries. The physical addresses discovered by
ARP can be provided to Data Link layer protocols. All addresses in
the ARP table are only local addresses. Any non-local address contains the hardware address of the local port on the router that is
used to access that non-local segment.
Internet Control Message Protocol (ICMP)
The Internet Control Message Protocol (ICMP) enhances the error
control provided by IP. Connectionless protocols, such as IP, cannot
detect internetwork errors, such as congestion or path failures.
ICMP can detect such errors and notify IP and upper-layer protocols. A network card that is generating an error often delivers a message to other network cards, via an ICMP packet.
TCP/IP Services
This section focuses on some of the TCP/IP services that exist within
the TCP/IP protocol suite. These services are just some of the more
common ones that you would deal with on a Microsoft network.
TRANSPORT PROTOCOL S
263
12.919-x CH07.i
264
8/28/98 1:10 PM
C h a p te r 7
Page 264
TR A NS P O R T P ROTOCOLS
Dynamic Host Configuration Protocol (DHCP)
When dealing with IP addressing, it can be very managementintensive to manually assign IP addresses and subnet masks to every
computer on the network. The Dynamic Host Configuration
Protocol (DHCP) enables automatic assignment of IP addresses.
This is usually performed by one or more computers (DHCP
Servers) that assigns IP addresses and subnet masks, along with other
configuration information, to a computer as it initializes on the network.
Most routers are configured not to forward broadcasts. DHCP, however, exchanges information by issuing broadcasts. A DHCP server,
therefore, needs to be on each segment. An alternative to placing a
DHCP server on each segment is to have a DHCP relay agent that
forwards on the client’s broadcast request for an IP address to a
DHCP server on another segment.
Domain Name System (DNS)
The Domain Name System (DNS) protocol provides host name and
IP address resolution as a service to client applications. DNS servers
enable humans to use logical node names, utilizing a fully qualified
domain name structure, to access network resources. Host names can
be up to 260 characters long.
Windows Internet Naming Services (WINS)
Windows Internet Naming Service (WINS) provides a function similar to that of DNS, with the exception that it provides NetBIOS
names to IP address resolution. This is important, because all of
Microsoft’s networking requires the ability to reference NetBIOS
names. Normally NetBIOS names are obtained with the issuance of
broadcasts, but because routers normally do not forward broadcasts,
a WINS server is one alternative that can be used to issue IP addresses to NetBIOS name requests.
File Transfer Protocol (FTP)
The File Transfer Protocol (FTP) is a protocol for sharing files
between networked hosts. FTP enables users to log on to remote
hosts. Logged-on users can inspect directories, manipulate files, execute commands, and perform other commands on the host. FTP
12.919-x CH07.i
8/28/98 1:10 PM
Page 265
C hapter 7
also has the capability of transferring files between dissimilar hosts
by supporting a file request structure that is independent of specific
operating systems.
Simple Mail Transfer Protocol (SMTP)
The Simple Mail Transfer Protocol (SMTP) is a protocol for routing
mail through internetworks. SMTP uses the TCP and IP protocols.
SNMP doesn’t provide a mail interface for the user. Creation, management, and delivery of messages to end users must be performed
by an email application.
Remote Terminal Emulation (TELNET)
TELNET is a terminal emulation protocol. TELNET enables PCs
and workstations to function as dumb terminals in sessions with
hosts on internetworks. TELNET implementations are available for
most end-user platforms, including UNIX (of course), DOS,
Windows, and Macintosh OS.
Network File System (NFS)
Network File System (NFS), developed by Sun Microsystems, is a
family of file-access protocols that are a considerable advancement
over FTP and TELNET. Because Sun made the NFS specifications
available for public use, NFS has achieved a high level of popularity.
NFS consists of two protocols:
á eXternal Data Representation (XDR). Supports encoding of data
in a machine-independent format. C programmers use XDR
library routines to describe data structures that are portable
between machine environments.
á Remote Procedure Call (RPC). Functions as a service request
redirector that determines whether function calls can be satisfied locally or must be redirected to a remote host. Calls to
remote hosts are packaged for network delivery and transmitted to RPC servers, which generally have the capability of servicing many remote service requests. RPC servers process the
service requests and generate response packets that are
returned to the service requester.
TRANSPORT PROTOCOL S
265
12.919-x CH07.i
266
8/28/98 1:10 PM
C h a p te r 7
Page 266
TR A NS P O R T P ROTOCOLS
TCP/IP Routing Protocols
The following sections describe two of the most common routing
protocols used by TCP/IP.
Routing Information Protocol (RIP)
The Routing Information Protocol (RIP) in the TCP/IP suite is not
the same protocol as RIP in the NetWare suite, although the two
serve similar functions. Internet RIP performs route discovery by
using a distance-vector method, calculating the number of hops that
must be crossed to route a packet by a particular path.
Although it works well in localized networks, RIP presents many weaknesses that limit its utility on wide-area internetworks. RIP’s distancevector route discovery method, for example, requires more broadcasts and thus causes more network traffic than some other methods.
The entire route table is also sent out on the broadcast, causing large
amounts of traffic as route tables become large. The Open Shortest
Path First (OSPF) protocol, which uses the link-state route discovery
method, is gradually replacing RIP. (See Chapter 6, “Connectivity
Devices and Transfer Mechanisms,” for more on routing.)
Open Shortest Path First (OSPF)
The Open Shortest Path First (OSPF) protocol is a link-state routediscovery protocol that is designed to overcome the limitations of
RIP. On large internetworks, OSPF can identify the internetwork
topology and improve performance by implementing load balancing
and class-of-service routing.
NetWare IPX/SPX
The protocols utilized with NetWare are summarized in Figure 7.4.
The NetWare protocols have been designed with a high degree of
modularity. This modularity makes the NetWare protocols adaptable
to different hardware and simplifies the task of incorporating other
protocols into the suite. Windows NT doesn’t use the IPX/SPX suite
to communicate with NetWare resources. Microsoft instead developed a clone of IPX/SPX called NWLink—IPX/SPX Compatible
Transport. IPX/SPX is generally smaller and faster than TCP/IP and,
like TCP/IP, it is routable. However, it operates down to the Data
Link layer of the OSI model so it is more dependent upon hardware
devices than the TCP/IP protocol.
12.919-x CH07.i
8/28/98 1:10 PM
Page 267
C hapter 7
TRANSPORT PROTOCOL S
FIGURE 7.4
OSI layer
The NetWare protocol architecture.
Application
NCP
SAP
Named Pipes
NetBIOS
Presentation
Session
SPX
Transport
RIP
IPX
NLSP
Network
LAN Drivers
Data Link
ODI
Physical
NDIS
Physical
General IPX/SPX Transport Protocols
The following subsections deal with protocols in the IPX/SPX protocol suite that relate back to the Network and Transport layers of
the OSI model.
Addressing in IPX
Addressing in IPX/SPX (NWLink) is much simpler than that in
TCP/IP. IPX/SPX also has two distinct addresses: a host address and
a network address. Unlike TCP/IP, the host address, or ID, is often
something that is not configured by the administrator.
The host address in IPX/SPX is based on the hardware address of
the network adapter card used by the device attaching to the network. These addresses are hexadecimal in nature, and address ranges
used by network adapter cards are assigned by the IEEE. Usually the
first two to three sets of numbers indicate the manufacturer of the
network adapter card.
Two examples of these addresses are:
44-45-53-54-00-00
07-00-4d-55-64-3e
As for the network address, this logical address is assigned by the
administrator of the cable segment. Usually when a server or router
267
12.919-x CH07.i
268
8/28/98 1:10 PM
C h a p te r 7
Page 268
TR A NS P O R T P ROTOCOLS
is installed, the logical network address is assigned by the administrator (see Exercise 7.2 for an example). The logical network address is
an eight-character hexadecimal address. Some possible examples
include:
903E0467
BEEF0000
E8012000
Again, any set of hexadecimal values is acceptable, but each network
address must be unique on the internetwork.
In general, addresses in an IPX/SPX network are often represented as
Host address : Network Address, as seen below:
55-66-00-e4-7a : E8022000
This address represents Host 55-66-00-e4-7a on Network
E8022000.
IPX
The Internetwork Packet Exchange Protocol (IPX) is a Network layer
protocol that provides connectionless (datagram) service. (IPX was
developed from the XNS protocol originated by Xerox.) As a
Network layer protocol, IPX is responsible for internetwork routing
and for maintaining network logical addresses. Routing uses the RIP
protocol (described later in this section) to make route selections.
IPX provides similar functionality as UDP does in the TCP/IP protocol suite.
IPX relies on hardware physical addresses found at lower layers to
provide network device addressing. IPX also uses sockets, or upperlayer service addresses, to deliver packets to their ultimate destinations. On the client, IPX support is provided as a component of the
older DOS shell and the current DOS NetWare requester. Windows
3.1 utilizes the DOS shell client, whereas Windows 95 and
Windows NT supports IPX if you install a Novell-supplied client.
Microsoft-supplied clients use the NWLink transport protocol supplied by Microsoft.
SPX
Sequenced Packet Exchange (SPX) is a Transport layer protocol that
extends IPX to provide connection-oriented service with reliable
12.919-x CH07.i
8/28/98 1:10 PM
Page 269
C hapter 7
delivery. Reliable delivery is ensured by the retransmittal of packets
in the event of an error. SPX is derived from a similar SPX protocol
in the XNS network protocol suite.
SPX establishes virtual circuits called connections. The connection ID
for each connection appears in the SPX header. A given upper-layer
process can be associated with multiple-connection IDs.
SPX is used in situations where reliable transmission of data is needed. SPX sequences the packets of data. Missing packets or packets
that don’t arrive in the order in which they were sent are detected
immediately. In addition, SPX offers connection multiplexing,
which is used in the printing environment. Many accounting programs, for example, call upon the services of SPX to ensure that data
is sent accurately. On the client, SPX support is provided as a component of the older DOS shell and of the current NetWare
requester. SPX provides functionality similar to that of TCP in the
TCP/IP protocol suite.
As a network administrator, you do not often get to pick whether
you wish to use IPX or SPX. It is often the applications one uses
that are preprogrammed to use one or the other. For example, in
most Novell networks, all file transfers are done using IPX. In the
case of printing, SPX is the protocol used.
Frame Type
When dealing with the IPX/SPX protocol suite, frame type is an
important issue. Frame type deals with the issue of how the data is
read by the adapter card. As you have seen in earlier chapters, data is
transmitted in digital format within a computer, and the network
card converts this digital information into a signal. This signal not
only contains the data being transferred, but also headers of information being used by all the protocols in the OSI seven layers.
When this data arrives at its destination, it gets converted from a
signal back into a recognizable format understood by the computer.
Frame type has to do with interpreting the bits of data as they come
in. As you will see in the following five sections, each of the five
frame types orders the information in the data differently than the
other frame types. Two computers not running the same frame type
cannot communicate.
When installing the IPX/SPX (or Microsoft’s NWLink) protocol on
a system, the frame type will either be automatically detected or
TRANSPORT PROTOCOL S
269
12.919-x CH07.i
270
8/28/98 1:10 PM
C h a p te r 7
Page 270
TR A NS P O R T P ROTOCOLS
must be manually assigned. Most modern computers can run multiple frame types at once.
The frame types to be discussed below include:
á 802.2
á 802.3
á Ethernet II
á Ethernet_SNAP
á Token-Ring
á Token-Ring_SNAP
802.2
The 802.2 frame type is the default frame type used on ethernet networks by all NetWare versions from 3.12 and onwards. What this
means is that this is the frame type these networking products use by
default. Figure 7.5 shows an initial breakdown of an 802.2 packet.
802.3
This was the default frame type used in all Novell NetWare products
from 3.11 and earlier. Figure 7.6 shows a breakdown of an 802.3
packet. The 802.3 packet is the same as an 802.2 packet, except that
the 802.3 packet does not contain the Destination Service Access
Point, Source Service Access Point, or Control bits.
Ethernet II
Ethernet II frame types are similar to 802.3 frame types, except they
contain a type field rather than a length field (see Figure 7.7). This
frame type can also be used with TCP/IP and AppleTalk.
Length 2bytes
Preamble
8bytes
FIGURE 7.5
An 802.2 frame type packet.
Destination
address
6bytes
Data 46-1500bytes
Frame Check
Sequence
4bytes
Source
address
6bytes
Destination Service Access Point
1byte
Source Service Access Point
1byte
Control
1byte
12.919-x CH07.i
8/28/98 1:10 PM
Page 271
C hapter 7
Length 2bytes
Preamble
8bytes
Destination
address
6bytes
Destination
address
6bytes
Frame Check
Sequence
4bytes
FIGURE 7.6
Data 46-1500bytes
Frame Check
Sequence
4bytes
FIGURE 7.7
Data 46-1500bytes
An 802.3 frame type packet.
Source
address
6bytes
Type 2bytes
Preamble
8bytes
TRANSPORT PROTOCOL S
An Ethernet II frame type packet.
Source
address
6bytes
(IPX/SPX);802.3>
Ethernet_SNAP
Ethernet_SNAP can be used for TCP/IP and AppleTalk Phase II
transport protocols, as well as IPX/SPX (see Figure 7.8).
Token-Ring
Token-ring frames are of two types. One is used to carry management information; the other is used to transfer data. Token-ring
frames are used on token-ring networks, and not ethernet networks.
Token-Ring_SNAP
A variation of the token-ring frame type is called token-ring_SNAP.
The token-ring_SNAP provides a function similar to that of the ethernet_SNAP frame type, but for token-ring networks.
Length 2bytes
Preamble
8bytes
Destination
address
6bytes
Frame Check
Sequence
4bytes
Data 46-1500bytes
Source
address
6bytes
Destination Service Access Point
1byte
Source Service Access Point
1byte
Control
1byte
Ethernet type 2bytes
Organization code 3bytes
FIGURE 7.8
An Ethernet_SNAP frame type packet.
271
12.919-x CH07.i
272
8/28/98 1:10 PM
C h a p te r 7
Page 272
TR A NS P O R T P ROTOCOLS
IPX/SPX Services
IPX/SPX services are similar to those used by TCP/IP in that they
provide a service to the user rather than being solely concerned with
transport issues. These services presented usually require the use of
either IPX or SPX as their transport mechanism, although recently
the capability to port these services over to TCP/IP has been included. Two services are briefly discussed in the following subsections.
These are SAP and NCP.
Service Advertising Protocol (SAP)
With Service Advertising Protocol (SAP), a device provides location
information by indicating what services it is offering. Devices can see
each other on the network by listing the SAPs each server issues. In
the case of NetWare, by default a SAP is issued every minute, telling
other computers what service the server is offering, as well as on
which node on what network this server is located.
NetWare Core Protocol (NCP)
The NetWare Core Protocol (NCP) provides numerous function calls
that support network services, such as file service, printing, name
management, file locking, and synchronization. NetWare client software interfaces with NCP to access NetWare services. NCP is to
NetWare networks as SMB is to Microsoft networks (see the section
titled “Server Messaging Blocks” later in this chapter).
NCP is a high-level protocol built into the NetWare operating system kernel. NCP covers aspects of the Session, Presentation, and
Application layers of the OSI reference model and has its own
miniature language that programmers use when writing applications
for the NetWare environment. The commands that NCP understands are associated primarily with access to files and directories on
a file server.
IPX/SPX Routing
This section looks at some of the more common routing protocols
that can be used in a network running IPX/SPX.
Router Information Protocol (RIP)
The Router Information Protocol (RIP) uses the distance vector route
discover method to determine hop counts to other devices. Like
12.919-x CH07.i
8/28/98 1:10 PM
Page 273
C hapter 7
IPX, RIP was developed from a similar protocol in the XNS protocol suite. RIP is implemented as an upper-layer service and is
assigned a socket (service address). RIP is based directly on IPX and
performs Network layer functions.
NetWare Link Services Protocol (NLSP)
NetWare Link Services Protocol (NLSP) is a link-state routing protocol used by routers (NetWare servers with two or more adapter cards
can act as routers) to advertise networks when their address tables
change.
NetBEUI
NetBEUI is a transport protocol that serves as an extension to
Microsoft’s Network Basic Input/Output System (NetBIOS).
Because NetBEUI was developed for an earlier generation of DOSbased PCs, it is small, easy to implement, and fast. It is actually the
fastest transport protocol available with Windows NT. Because it was
built for small, isolated LANs, however, NetBEUI is non-routable,
making it somewhat anachronistic in today’s diverse and interconnected networking environment. NetBEUI is also a broadcast-based
protocol and as such can cause congestion in larger networks.
Fortunately, the NDIS standard enables NetBEUI to coexist with
other routable protocols. For instance, you could use NetBEUI for
fast, efficient communications on the LAN segment and use TCP/IP
for transmissions that require routing (see Exercise 7.2).
AppleTalk
AppleTalk is the computing architecture developed by Apple
Computer for the Macintosh family of personal computers.
Although AppleTalk originally supported only Apple’s proprietary
LocalTalk cabling system, the suite has been expanded to incorporate both ethernet and token-ring Physical layers. Within Microsoft
operating systems, AppleTalk is only supported by Windows NT
Server. Windows NT Workstation and Windows 95 do not support
AppleTalk. AppleTalk cannot be used for Microsoft-to-Microsoft
operating system communication. It can be used only through
Windows NT servers supporting Apple clients.
TRANSPORT PROTOCOL S
273
12.919-x CH07.i
274
8/28/98 1:10 PM
C h a p te r 7
Page 274
TR A NS P O R T P ROTOCOLS
AppleTalk originally supported networks of limited scope. The
AppleTalk Phase II specification issued in 1989, however, extended
the scope of AppleTalk to enterprise networks. The Phase II specification also enabled AppleTalk to coexist on networks with other protocol suites. Figure 7.9 presents a layered perspective of the
AppleTalk protocols.
The LocalTalk, EtherTalk, and TokenTalk Link Access Protocols
(LLAP, ELAP, and TLAP) integrate AppleTalk upper-layer protocols
with the LocalTalk, ethernet, and token-ring environments.
Apple’s Datagram Deliver Protocol (DDP) is a Network layer protocol
that provides connectionless service between two sockets. A socket is
the AppleTalk term for a service address. A combination of a device
address, network address, and socket uniquely identifies each
process.
DDP performs network routing and consults routing tables maintained by Routing Table Maintenance Protocol (RTMP) to determine routing. Packet delivery is performed by the data link protocol
operating on a given destination network.
The AppleTalk Transaction Protocol (ATP) is a connectionless
Transport layer protocol. Reliable service is provided through a system of acknowledgments and retransmissions. Retransmissions are
initiated automatically if an acknowledgment is not received within a
specified time interval. ATP reliability is based on transactions. A
transaction consists of a request followed by a reply. ATP is responsible for segment development and performs fragmentation and
reassembly of packets that exceed the specifications for lower-layer
protocols. Packets include sequence numbers that enable message
reassembly and retransmission of lost packets. Only damaged or lost
packets are retransmitted.
Application
AppleShare
AFP
Presentation
Session
The AppleTalk protocol suite.
ZIP PAP
ATP
Network
DDP
Data Link
FIGURE 7.9
ADSP
Transport
Physical
LocalTalk
ASP
NBP
RTMP
AARP
EtherTalk
TokenTalk
12.919-x CH07.i
8/28/98 1:10 PM
Page 275
C hapter 7
The AppleTalk File Protocol (AFP) provides file services and is
responsible for translating local file service requests into formats
required for network file services. AFP directly translates command
syntax and enables applications to perform file format translations.
AFP is responsible for file system security, and verifies and encrypts
logon names and passwords during connection setup.
AppleShare is a client/server system for Macintosh. AppleShare provides three primary application services:
á The AppleShare File Server uses AFP to enable users to store
and access files on the network. It logs in users and associates
them with network volumes and directories.
á The AppleShare Print Server uses NBP and PAP to support
network printing. NBP provides name and address information that enables PAP to connect to printers. The AppleShare
Print Server performs print spooling and manages printing on
networked printers.
á The AppleShare PC enables PCs running MS-DOS to access
AppleShare services by running an AppleShare PC program.
Data Link Control (DLC)
The Data Link Control (DLC) protocol does not provide a fullyfunctioning protocol stack. In Windows NT systems, DLC is used
primarily to access Hewlett-Packard JetDirect network-interface
printers. DLC also provides some connectivity with IBM mainframes and for the Windows NT remoteboot service used by
Diskless Windows 95 workstations. DLC is not a protocol that can
be used to connect Windows NT or 95 computers together.
RELATING PROTOCOL STACKS
TOGETHER
Discussing protocol stacks can be a very daunting exercise because
many different protocols exist at every layer in the OSI model and
make up the entire stack.
TRANSPORT PROTOCOL S
275
12.919-x CH07.i
276
8/28/98 1:10 PM
C h a p te r 7
Page 276
TR A NS P O R T P ROTOCOLS
As was described during the discussion of TCP/IP, IPX/SPX, and
NetBEUI, these protocol stacks are separated into three areas (except
for NetBEUI). These areas are:
á Transport protocols
á Services offered
á Routing protocols
The transport protocols are the rules or standards used by the protocol stack to facilitate the movement of data between different
devices. The services offered are some of the unique network services
offered by each protocol stack. The routing protocols are designed to
enable dynamic routing. The actual address format is based upon
either TCP/IP (four octets of numbers) or IPX/SPX (network
address was eight-character hexadecimal).
The interesting aspect of protocol stacks is the services they offer.
Take TCP/IP for example. TCP/IP was developed to connect dissimilar machines together, and to provide services that could be used by
these interconnected machines.
NOTE
Another example would be that of FTP. This service was developed
to transfer files from one machine to another. Thus an FTP client
and an FTP server service were developed. The specification between
these two programs was made public. If you programmed your FTP
client to the open standards, it would interoperate with the FTP service that was also programmed to the open standards. Both the
client and the service were designed to only “hook into” or operate
with the TCP and IP Transport and Network layers of the protocol
stack. That is, the FTP client or service is not programmed to operate with the NetBEUI transport protocol, for example.
NetWare and TCP/IP Add-ons to
NetWare enable it to utilize TCP/IP,
and the new NetWare version 5 is
also being designed to use TCP/IP.
The same analogy can be applied to NCP. NCP packets are designed
to communicate with a NetWare server. Novell developed the NCP
standard, and this standard is designed to operate only with IPX;
thus communication to a Novell NetWare server cannot be done
over NetBEUI or TCP/IP, because the NCP protocol is not designed
to interoperate with these transport protocols.
Server Messaging Blocks (SMB)
One protocol that is slightly independent is Microsoft’s Server
Messaging Blocks (SMB). SMB’s are Microsoft’s equivalent to NCP
12.919-x CH07.i
8/28/98 1:10 PM
Page 277
C hapter 7
packets. Like NCP packets, SMB’s operate at the Application layer
of the OSI model.
SMBs enable machines on a Microsoft network to communicate with
one another. Through the use of SMBs, File and Print services can
be shared. SMBs can use TCP/IP, NWLink (IPX/SPX), or NetBEUI,
because SMBs utilize a NetBIOS interface when communicating.
For more information on NetBIOS names, see the following section.
NETBIOS NAMES
Implement a NetBIOS naming scheme for all computers on a given
network.
NetBIOS is an interface that provides NetBIOS-based applications
with access to network resources. Every computer on a Windows
NT network must have a unique name for it to be accessible
through the NetBIOS interface. This unique name is called a computer name or a NetBIOS name.
NetBIOS Background
NetBIOS (Network Basic Input/Output System) is an application
interface that provides PC-based applications with uniform access to
lower protocol layers. NetBIOS was once most closely associated
with the NetBEUI protocol—NetBEUI, in fact, is an abbreviation
for NetBIOS Extended User Interface. In recent years, however,
other vendors have recognized the importance of providing compatibility with PC-based applications through NetBIOS, and NetBIOS
is now available with many protocol configurations. For instance,
such terms as “NetBIOS over IPX” or “NetBIOS over TCP/IP” refer
to the protocols used with NetBIOS.
All of Microsoft’s networking architecture references NetBIOS
names, thus it can be said that Microsoft Networking uses a
NetBIOS interface to access components on the network.
Assigning NetBIOS Names
On a NetBIOS network, every computer must have a unique name.
The computer name must be 15 characters long or fewer. A
TRANSPORT PROTOCOL S
277
12.919-x CH07.i
278
8/28/98 1:10 PM
C h a p te r 7
Page 278
TR A NS P O R T P ROTOCOLS
NetBIOS name can include alphanumeric characters and any of the
following special characters:
[email protected]#$%^&()-_’{}.~
Note that you cannot use an asterisk or all periods in a NetBIOS
name. It is also not recommended to use spaces in NetBIOS names
as well, as some applications are not able to work with a space in a
NetBIOS name. Also, NetBIOS names are not case-sensitive.
Within these character limitations, you can choose any name for a
PC. The rule of thumb is to choose a name that helps you to identify the computer. Names such as PC1, PC2, and PC3 are difficult to
visualize and easy to confuse. Likewise, names such as MYPC or
WORTHLESSPC could confuse you in the long run, especially if
you have many computers on your network. For these reasons,
names that include a hook relating the name of the owner or the
location of the computer generally are more effective. Consider the
following names, for example:
á BILLS_PC
á MARKETINGPC
á LUNCHROOM_PC
á BILLS_LAPTOP
You must specify a computer name for a Windows NT or Windows
95 computer at installation. The computer name then becomes part
of the network configuration. In either Windows NT or Windows
95, you can change the name of the computer through the Control
Panel Network application. (See the following Review Break and
Step-by-Steps 7.1 and 7.2.)
R E V I E W
B R E A K
A NetBIOS computer name must:
á Be unique
á Consist of 15 or fewer characters
á Consist of either alphanumeric characters or the characters
[email protected]#$%^&()-_’{}.~
12.919-x CH07.i
8/28/98 1:10 PM
Page 279
STEP BY STEP
7.1 Changing a Computer’s Name in Windows NT
You designate a computer name for your PC when you install the
operating system. You can change the computer name later through
the Control Panel Network application, but you must have
Administrative privileges on a Windows NT computer to change the
computer name. Anyone can change the NetBIOS name on a
Windows 95 computer. To change a NetBIOS computer name, follow these steps:
WA R N I N G
C hapter 7
TRANSPORT PROTOCOL S
279
Changing Names of Windows NT
Computers in a Domain Changing
the names of Windows NT computers that are in a domain can have
serious consequences. Problems
such as authentication not working
and applications not functioning
can occur. Consult an administrator
before doing this.
1. Click the Start button and choose Settings, Control Panel.
2. In Windows NT Control Panel, double-click the Network
application.
3. In the Network application’s Identification tab, click on
the Change button. The subsequent Identification
Changes dialog box is shown in Figure 7.10.
4. Change the computer name in the text box labeled
Computer Name and click OK.
As with Windows NT, Windows 95 enables you to change the computer name after installation by using the Control Panel Network
application. To change the name, follow these steps:
STEP BY STEP
7.2 Changing the Computer Name in Windows 95
1. Click the Start button and choose Settings, Control Panel.
2. In the Windows 95 Control Panel, double-click the
Network application.
3. In the Network application, choose the Identification tab.
4. To change the computer name, edit the text in the
Computer name text box (see Figure 7.11).
FIGURE 7.10
Windows NT’s Identification Changes dialog box
enables you to change the computer name.
12.919-x CH07.i
280
8/28/98 1:10 PM
C h a p te r 7
Page 280
TR A NS P O R T P ROTOCOLS
A computer on a NetBIOS network must have a NetBIOS computer name. The NetBIOS name is configured at installation and, in
Windows NT or Windows 95, can be changed later through the
Control Panel Network application. Computers use the NetBIOS
name (sometimes combined with a share name or a path name) to
locate resources on the network.
Finding Resources on Microsoft
Networks
The Universal Naming Convention is a standard for identifying
resources on Microsoft networks. A UNC path consists of the following components:
á A NetBIOS computer name preceded with a double backslash
(left-leaning slash)
FIGURE 7.11
The Identification tab of the Windows 95
Network dialog box enables you to change the
computer name.
á The share name of a shared resource located on the given PC
(optional—see Chapter 10 for more details concerning shares)
á The MS-DOS–style path of a file or a directory located on the
given share (optional)
Elements of the UNC path are separated with single backslashes.
The following list details some legal UNC names:
\\BILL’s_PC
\\WEIGHTRM\ACCOUNTS
\\PET_DEPT\CATS\SIAMESE.TXT
Various Windows NT commands use UNC paths to designate network resources. For instance,
net view \\PET_DEPT
enables you to view the shared resources on the computer with the
NetBIOS name PET_DEPT. The command
net use G: \\PET_DEPT\CATS
maps the shared directory CATS on the computer PET_DEPT to
drive G:.
12.919-x CH07.i
8/28/98 1:10 PM
Page 281
C hapter 7
TRANSPORT PROTOCOL S
281
C A S E S T U DY : S E L E C T I N G P R O T O C O L S TA C K S
ESSENCE OF THE CASE
A N A LY S I S
The essence of the case is as follows:
Basically, five protocol stacks are supported by
Windows NT Server and four by Windows 95.
TCP/IP, NWLink (IPX/SPX), NetBEUI, and DLC are
supported by both operating systems, and
AppleTalk is supported by Windows NT Server.
• Summarize the protocol stacks
• Address the benefits of the protocol
stack
• Address the issues with the protocol
stack
• Address the possibility of connecting to
the Internet in the future
• Address the possibility of Novell servers
in the future
SCENARIO
You have been invited into a network planning
session. A new corporate-wide internetwork is
being rolled out. The network will consist of
Microsoft Windows NT Servers and Windows 95
workstations. The only service you are aware will
be running on the internetwork is File and Print
services. All computers need to be able to connect with all other computers on the internetwork, thus a routable protocol needs to be used.
Possible future changes to the network include
connection to the Internet or the addition of
some Novell servers. You are to make a summary of which protocol stacks should be used on
the network. Basically you are to recommend a
type of protocol, describe why it should be implemented, and explain the benefits and any concerns about using this protocol.
If Novell servers are to be added to the internetwork in the future, NWLink should be used. If
they will connect to the Internet, TCP/IP may
need to be used.
The five protocols are analyzed in the following
sections.
DLC
This protocol is not an option to connect the
computers together. This protocol would only be
added if there was to be some IBM mainframe
connectivity or if there were some HewlettPackard JetDirect printers to be managed.
Because neither of these options were specified,
there is no reason to install this protocol on the
machines.
AppleTalk
AppleTalk is only supported by Windows NT
Server, and only for communications between
Windows NT Server and Apple clients. Because
this protocol cannot be used for communication
between Microsoft machines, and no Apple computers are on the network, there is no reason to
install this protocol stack.
continues
12.919-x CH07.i
282
8/28/98 1:10 PM
C h a p te r 7
Page 282
TR A NS P O R T P ROTOCOLS
C A S E S T U DY : S E L E C T I N G P R O T O C O L S TA C K S
continued
NetBEUI
NetBEUI is a protocol that is supported by all the
Microsoft operating systems. It is a fast protocol
that really requires no management to maintain.
The only downfall is that NetBEUI is a nonroutable protocol. An additional protocol would be
needed for all LAN segments to intercommunicate. (As a side note, the networked game of
Hearts cannot function unless the NetBEUI protocol is installed.)
TCP/IP
The TCP/IP protocol stack meets several criteria.
It is a routable protocol and is the protocol needed to go out on the Internet. It does not, however, allow access to the Novell NetWare servers if
they are added in the future.
TCP/IP is a management-intensive protocol. To
ease the management issues, planning must be
done. If the network is supposed to connect to
the Internet, unique IP addresses must be
obtained from an Internet Service Provider (ISP).
Also, careful programming of the routers is necessary for internetwork connectivity to work.
Another issue that may need to be addressed is
whether the network will use DHCP to allocate IP
addresses on the network. If this is the case, a
DHCP server must be placed on each subnet,
unless a relay agent is placed on the subnet to
forward requests to a DHCP server on a remote
subnet.
Another issue would be the use of a WINS server. Microsoft networking uses NetBIOS names to
communicate all networking functions. NetBIOS
names need to be resolved to IP addresses so
that the transport stack of TCP/IP can move the
networking communication. NetBIOS names can
be resolved on a local subnet through the use of
broadcasts, but remote NetBIOS names cannot
be resolved. By using a WINS server, this issue
can be centrally managed.
NWLink (IPX/SPX)
Using NWLink enables the Windows NT and 95
machines to communicate with each other and
across the routers. It also enables the computers to talk to any Novell servers that may be
installed on the network. It does not, however,
enable communication with the Internet.
An issue to consider with NWLink is what frame
type is to be used on the network. Some equipment cannot handle certain frame types, thus it
would be important to get a list of all hardware
being used on the network to see whether it is
compliant with the frame type to be chosen. It
would be a good recommendation to choose the
802.2 frame type, because this is the current
IEEE recommended standard.
Conclusion
There is no compellingly correct answer. It looks
as if NWLink and TCP/IP are both contenders.
Until the company makes a decision to either go
onto the Internet and/or use Novell servers, a
single protocol cannot be favored over another.
12.919-x CH07.i
8/28/98 1:10 PM
Page 283
C hapter 7
C A S E S T U DY : I M P L E M E N T I N G
A
ESSENCE OF THE CASE
The essence of the case is as follows:
• You are to assign NetBIOS names to all
computers on the network.
• You do not need to follow any predetermined conventions.
SCENARIO
You are to implement a NetBIOS naming convention for the network. You are free to use what
names you feel are suitable.
A N A LY S I S
Being able to freely assign NetBIOS names can
be a pleasure and a pain. The pleasure is that
you are able to leave your own personal mark on
the network. The naming convention that you
choose is the one that others follow. The pain is
that other network administrators may criticize
your decisions.
When naming computers with NetBIOS names,
recall that only three rules need to be followed:
• All names must be unique.
• Names cannot be more than 15 characters
long.
TRANSPORT PROTOCOL S
283
N E T BIOS N A M I N G C O N V E N T I O N
• Alpha-numeric characters and the characters [email protected]#$%^&()-_{}.~ are allowed.
There are no right or wrong names (unless you
break the rules for characters). There are, however, efficient and inefficient names. Efficient
names enable an administrator or support person to identify the location of the computer
easily on the network. An inefficient name does
not enable one to locate the computer on the
network. Try to be descriptive and use something
that follows corporate naming traditions.
FS1, FS1, APPS_SERV, and PRINT_SERV are all
valid names that could be used for servers. All
your servers should be locked up in a server
room, so the name does not necessarily need to
describe the location, but instead be descriptive
of the function of the server.
F1PC3, ACCTDEPT#5 and ALVINSPC all are
descriptive in location. F1PC3 could describe that
this is the third PC on the first floor. ACCTROOM#5 could mean the fifth PC in the
Accounting department. ALVINSPC of course
means Alvin’s PC. Again, all PC names ought to
be descriptive so that the unique identifier also
provides useful information such as location
and/or function.
12.919-x CH07.i
284
8/28/98 1:10 PM
C h a p te r 7
Page 284
TR A NS P O R T P ROTOCOLS
CHAPTER SUMMARY
KEY TERMS
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Packet
Protocol
TCP/IP
Internet Protocol (IP)
Transmission Control Protocol (TCP)
User Datagram Protocol (UDP)
Address Resolution Protocol (ARP)
Internet Control Message Protocol
(ICMP)
Dynamic Host Configuration Protocol
(DHCP)
Windows Internet Naming Service
(WINS)
File Transfer Protocol (FTP)
Simple Mail Transport Protocol (SMTP)
Remote Terminal Emulation (Telnet)
Network File System (NFS)
Routing Information Protocol (RIP)
Open Shortest Path First (OSPF)
IPX/SPX
NWLink
Internetwork Packet Exchange (IPX)
Sequenced Packet Exchange (SPX)
802.2 frame
802.3 frame
Ethernet II frame type
Ethernet_SNAP frame
Token-ring frame
Token-ring_SNAP frame
Service Advertising Protocol (SAP)
NetWare Core Protocol (NCP)
NetWare Link Services Protocol (NLSP)
NetBEUI
AppleTalk
Data Link Control (DLC)
Server Messaging Blocks (SMB)
This chapter examined network protocols and protocol suites. The
chapter began with an introduction to protocol stacks. You then
learned about some of the most common protocol suites, as follows:
á TCP/IP. The Internet protocol suite
á IPX/SPX. A protocol suite used for Novell NetWare networks
á NetBEUI. A non-routable protocol used on Microsoft net-
works
á AppleTalk. The Apple Macintosh protocol system
á DLC. A protocol that Windows NT networks use to connect
with HP JetDirect printers and IBM mainframes
During the discussion on the protocol suites, analysis was focused
on the addressing schemes used by each suite, as well as the smaller
components and their functions within each protocol suite.
The NDIS interface standard (discussed in Chapter 2) enables a single computer to bind one network adapter to more than one protocol system. This provides great versatility and interoperability in
today’s diverse networking environment.
The chapter also discussed NetBIOS names and naming conventions. You saw that all Microsoft machines on a network use
NetBIOS names and that these names have some rules regarding
their construction.
12.919-x CH07.i
8/28/98 1:10 PM
Page 285
C hapter 7
TRANSPORT PROTOCOL S
285
A P P LY Y O U R L E A R N I N G
Exercises
7.1
Installing Network Protocols in
Windows NT
Objective: Become familiar with the procedure for
installing and removing protocols in Windows NT.
Estimated Time: 15 minutes
1. You can install, configure, remove, and manage
network protocols by using the Network application in Windows NT’s Control Panel. Click the
Start menu and choose the Settings/Control panel.
Then double-click the Network application icon.
Another way to reach the Network application is
to right-click the Network Neighborhood icon
and choose Properties.
2. In the Network application, choose the Protocols
tab (see Figure 7.12). The Network Protocols box
displays the protocols currently installed on the
system.
3. If TCP/IP is installed on your system, select
TCP/IP Protocol and choose Properties to invoke
the Microsoft TCP/IP Properties dialog box (see
Figure 7.13). Note the several tabs that provide
various configuration options. Close the
Microsoft TCP/IP Properties dialog box and
select the NetBEUI protocol (if it is installed) in
the Network application’s Protocols dialog box.
Note that the Properties button is grayed. Try
double-clicking the NetBEUI icon in the box’s
list of protocols. A message says Cannot configure
the software component. Unlike TCP/IP,
NetBEUI is not user-configurable.
If TCP/IP and NetBEUI aren’t installed on
your system, you can install them by using the
procedure described in steps 4 and 5 of this exercise and then delete them later.
4. To add a protocol, click on the Add button in the
Network application’s Protocols tab. Select a protocol from the protocol list in the Select Network
FIGURE 7.12
FIGURE 7.13
Network application’s Protocols tab.
The Microsoft TCP/IP Properties dialog box.
12.919-x CH07.i
286
8/28/98 1:10 PM
C h a p te r 7
Page 286
TR A NS P O R T P ROTOCOLS
A P P LY Y O U R L E A R N I N G
Protocol dialog box (see Figure 7.14). Click on
OK to install the protocol. Windows NT may
prompt you for the location of the Windows NT
installation disk. Type in the location or drive letter of the installation files and click OK. If you
are installing a protocol that requires some configuration (such as TCP/IP or NWLink), Windows NT asks you for the necessary information.
5. Windows NT asks you to restart your system.
Shut down your system and restart. Return to the
Network application’s Protocols tab and see
whether the protocol is properly installed.
6. To remove a protocol, select the protocol from
the Network Protocols list and click on the
Remove button.
between protocol layers that enables those layers to
behave like a protocol stack. By binding a transport
protocol such as TCP/IP (which operates at the
Transport and Network levels) to a network adapter
(which operates at the Data Link and Physical layers)
you provide a conduit for the protocol’s packets to
reach the network and thus enable the protocol to participate in network communications. NDIS lets you
bind multiple protocols to a single adapter or multiple
adapters to a single protocol.
1. Click the Start button and choose Settings,
Control Panel. In Windows NT’s Control Panel,
double-click the Network application icon and
choose the Bindings tab (see Figure 7.15).
2. Click the Show Bindings down arrow to access
the drop-down list. Note that you can display
bindings for services, protocols, or adapters. A
service bound to a protocol bound to an adapter
provides a complete pathway from the local system to the network.
FIGURE 7.14
The Select Network Protocol dialog box.
7.2
Network Bindings
Objective: Become familiar with the process for
enabling and disabling network bindings and changing
network access order.
Estimated Time: 10 minutes
In Chapter 2, you learned about NDIS and the concept of network bindings. A binding is an association
FIGURE 7.15
Network application’s Bindings tab.
12.919-x CH07.i
8/28/98 1:10 PM
Page 287
C hapter 7
TRANSPORT PROTOCOL S
287
A P P LY Y O U R L E A R N I N G
3. Click on the plus sign next to the Workstation
service. The Workstation service is the Windows
NT redirector (refer to Chapter 1, “Networking
Terms and Concepts”), which redirects requests
from the local system to the network. The protocols currently bound to the Workstation service
appear in a list below the Workstation icon. Click
on the plus sign next to one of the protocols. The
network adapters bound to the protocol now
appear in the tree (see Figure 7.16).
4. The protocols and their associated adapters represent potential pathways for the Workstation
service to access the network. Windows NT prioritizes those pathways according to the order in
which they appear in the Bindings tab. For the
configuration shown in Figure 7.16, for example,
Windows NT attempts to use the NetBEUI protocol with the Workstation service before
attempting to use NWLink. The Move Up and
Move Down buttons let you change the access
order. Select a protocol under the Workstation
service. Try the Move Up and Move Down buttons to change the position of the protocol in the
access order. (Don’t forget to restore the protocol
to its original position before leaving the
Bindings tab.)
5. The Enable and Disable buttons let you enable or
disable a protocol for a given service, because you
may not wish people using a particular protocol
to use a certain service. Disable a protocol (for
instance, NetBEUI) for the Workstation service.
Now click the plus sign next to the Server service.
Note that although the protocol is disabled for
the Workstation service, it is still enabled for the
Server service. Re-enable the protocol under the
Workstation service and close the Network application.
7.3
Mapping a Network Drive
Objective: Use the NetBIOS-based UNC path to map
a drive letter to a network share.
Estimated Time: 10 minutes
1. Double-click Windows NT’s Network
Neighborhood application. Locate another computer for which network shares have been
defined.
Another useful tool for finding network shares is
the Server Manager application in Windows NT
Server’s Administrative Tools group. To use this
tool, click the Start menu and choose Programs,
Administrative Tools, Server Manager.
2. Click the Start menu and go to the Windows NT
command prompt. (Choose Programs,
Command Prompt.)
FIGURE 7.16
Inspect binding information by using the Bindings tab.
3. Enter the following command:
net view
12.919-x CH07.i
288
8/28/98 1:10 PM
C h a p te r 7
Page 288
TR A NS P O R T P ROTOCOLS
4. The net view command lists the NetBIOS names
of computers in your domain. Look for the computer you located using Network Neighborhood
in Step 1.
NOTE
A P P LY Y O U R L E A R N I N G
Privileges You must have the necessary privileges to access the shared
directory. Check with your network
administrator for details.
5. Type the following command:
net view \\computername
where computername is the NetBIOS name of
the computer you located in Step 1. This command lists the network shares available on the
computer.
6. Locate a directory share in the share list. Then
type the following command:
net use * \\computername\sharename
where computername is the NetBIOS name of
the computer you located in Step 1, and sharename is the name of the share you located in the
preceding step. The asterisk maps the next available drive letter to the share. You could also specify a particular drive letter (followed by a colon)
rather than the asterisk. A message appears on
your screen giving you the drive letter that
Windows NT used for the connection and indicating whether the command was successful.
9. Type the command dir and press the Enter key.
A directory listing of the shared directory should
appear on your screen. You now have accessed the
shared directory through the mapped drive letter.
10. To delete the network drive mapping, enter the
following command:
net use drive_letter: /delete
where drive_letter is the drive letter assigned in
Step 6.
You also can map drive letters through Windows NT
Explorer. To do so, pull down the Tools menu and
select Map Network Drive. In the Drive property box,
select the drive letter you wish to use. In the Path box,
type in \\Computer_Name\Share_Name, where
Computer_Name is the NetBIOS name of the computer to which you are connecting, and Share_Name is the
name of the shared directory on the other computer.
7. Now enter the following command:
net view \\computername
where computername is the name of the computer you chose in Step 1. The drive letter you
mapped to the share should appear beside the
share name, and the share type in the column
should be titled Used as.
8. Enter the drive letter assigned in Step 6 at the
command prompt, followed by a colon. For
instance, enter I:.
Review Questions
1. Name three transport protocols that can be used
to transfer SMBs and the only transport protocol
that can transfer NCP packets.
2. Which protocol suite cannot be used for PC-toPC communication when Microsoft operating
systems are running on both PCs?
3. How many characters can you have in a
NetBIOS name?
12.919-x CH07.i
8/28/98 1:10 PM
Page 289
C hapter 7
TRANSPORT PROTOCOL S
289
A P P LY Y O U R L E A R N I N G
Exam Questions
1. Which three of the following are Transport layer
protocols?
A. ALP
B. PPX
C. TCP
D. SPX
2. Which three of the following operate within the
Network layer?
A. Telnet
B. WINS
A. NWLink
B. TCP/IP
C. DLC
D. NetBEUI
6. What is the best protocol for a remote PC that
interacts with the network via the Internet?
A. NWLink
B. TCP/IP
C. DLC
D. NetBEUI
7. At which OSI layer does NCP operate?
C. FTP
A. Application and Presentation
D. IP
B. Transport and Network
3. At which OSI layer does SMB operates?
C. Network only
A. Application
D. Transport only
B. Transport
E. Session and Transport
C. Network
D. Physical
4. Which three of the following protocols are available with Windows NT?
A. AppleTalk
B. IPX/SPX
C. NetBEUI
D. DLC
5. What is the simplest protocol to use for an
isolated LAN with several DOS-based clients
and Windows NT Server?
8. UDP is part of which protocol suite?
A. TCP/IP
B. IPX/SPX
C. AppleTalk
D. NetBEUI
9. How does TCP/IP compare to NetBEUI on a
small network?
A. Faster
B. Slower
C. Easier to install and configure
D. None of the above
12.919-x CH07.i
290
8/28/98 1:10 PM
C h a p te r 7
Page 290
TR A NS P O R T P ROTOCOLS
A P P LY Y O U R L E A R N I N G
10. NetBIOS is an abbreviation for what?
A. Network Basic Input/Output System
B. Network Bilateral Operating System
14. Which 802 category defines the Ethernet
Standard?
A. 802.10
B. 802.3
C. Network Binary Interchange Operating
System
C. 802.12
D. Network Bus Input/Output System
D. 802.5
11. Which of the following is a legal and recommended NetBIOS computer name?
A. EAGLES_LODGE_PENT
B. [email protected]#*_PC
C. 486!!EAGLES_PC
D. EAGLES LODGE
12. Which of the following UNC paths lead you to a
file called DOUGHNUTS on a PC called
FOOD located in the SWEETS directory of the
JUNKFOOD share?
A. \\DOUGHNUTS\FOOD\SWEETS\JUNKFOOD
B. \\FOOD\JUNKFOOD\SWEETS\DOUGHNUTS
C. \\FOOD\JUNKFOOD\DOUGHNUTS
D. \\JUNKFOOD\DOUGHNUTS
13. Which of the following commands produces a
list of shared resources on the computer described
in Question 12?
A. Net share \\FOOD
B. Net view
C. Net view \\FOOD
D. Net view \\FOOD /shares
15. You wish to connect a Windows NT computer to
a Windows 95 computer to share files back and
forth.
Primary objective: You need to establish network
connectivity by using a compatible protocol.
Secondary objective: The Windows 95 computer
needs to connect directly to the Internet.
Secondary objective: You wish to administer HP
JetDirect printers from the Windows NT computer.
Suggested Solution: You install DLC and
NetBEUI on the Windows NT computer. You
install NetBEUI, DLC, NWLINK, and TCP/IP
on the Windows 95 computers. Make sure that
the NetBIOS names on both the Windows NT
and Windows 95 computers are the same.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objective.
D. This solution does not meet the primary
objective.
16. You wish to connect a Windows NT computer to
a Windows 95 computer to share files back and
forth.
12.919-x CH07.i
8/28/98 1:10 PM
Page 291
C hapter 7
TRANSPORT PROTOCOL S
291
A P P LY Y O U R L E A R N I N G
Primary objective: You need to establish network
connectivity by using a compatible protocol.
Secondary objective: The Windows 95 computer
needs to connect directly to the Internet.
Secondary objective: You wish to administer HP
JetDirect printers from the Windows NT computer.
Suggested Solution: You install DLC and
NetBEUI on the Windows NT computer. You
install DLC, NWLINK, and TCP/IP on the
Windows 95 computers. Make sure that the
NetBIOS names on both the Windows NT and
Windows 95 computers are different, and conform to the NetBIOS naming rules and recommendations.
Suggested Solution: You install DLC and
NetBEUI on the Windows NT computer. You
install NetBEUI, DLC, and NWLINK on the
Windows 95 computers. Make sure that the
NetBIOS names on both the Windows NT and
Windows 95 computers are different, and conform to the NetBIOS naming rules and recommendations.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objective.
D. This solution does not meet the primary
objective.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objective.
D. This solution does not satisfy the primary
objective.
17. You wish to connect a Windows NT computer to
a Windows 95 computer to share files back and
forth.
Primary objective: You need to establish network
connectivity by using a compatible protocol.
Secondary objective: The Windows 95 computer
needs to connect directly to the Internet.
Secondary objective: You wish to administer HP
JetDirect printers from the Windows NT computer.
Answers to Review Questions
1. SMBs are transferred between computers using
NWLink (IPX/SPX), TCP/IP, and NetBEUI.
DLC and AppleTalk cannot be used for SMB
transport.
NCP packets can only be transported using
IPX/SPX (NWLink). See the sections titled
"Server Message Blocks (SMB)" and "IPX/SPX."
2. DLC and AppleTalk cannot be used for PC-toPC communication when both PCs' operating
systems are Microsoft products. See the sections
titled "Data Link Control (DLC)" and
"AppleTalk."
3. A NetBIOS name can have up to 15 characters.
See the section titled "NetBIOS Names."
12.919-x CH07.i
292
8/28/98 1:10 PM
C h a p te r 7
Page 292
TR A NS P O R T P ROTOCOLS
A P P LY Y O U R L E A R N I N G
Answers to Exam Questions
1. C, D. Answers A and B do not exist. See the section titled "Transport Protocols." TCP is the
TCP/IP connection-oriented transport protocol,
whereas SPX is the IPX/SPX connection-oriented
transport protocol.
2. D. TCP operates only in the Transport layer.
Telnet, WINS, and FTP all operate at the layers
above the Transport layer. See the section titled
"Transmission Control Protocol" within the section titled "Transport Protocols."
7. A. NCP is an Application layer protocol used to
communicate with Novell servers. See the section
titled "NetWare Core Protocol" under the section
titled "NetWare IPX/SPX."
8. A. UDP operates at the Transport layer of the
TCP/IP protocol suite. See the section titled
"User Datagram Protocol (UDP)" under the section titled "TCP/IP—The Internet Protocols."
9. B. TCP/IP is slower than NetBEUI. It is harder
to configure because it requires knowledge of IP
addresses. See the section titled "NetBEUI."
3. A. SMB is an Application layer protocol. See the
section titled "Server Message Blocks."
10. A. All the other answers are made-up answers.
See the section titled "NetBIOS Names."
4. A, C, D. B is incorrect because Windows NT
does not have true IPX/SPX, but instead a version of IPX/SPX known as NWLink. See the section titled "NetWare IPX/SPX."
11. C. A is more than 15 characters, B contains a *,
and D has a space, which is not recommended.
See the section titled "NetBIOS Names."
5. D. D is correct because NetBEUI has really no
management considerations. A and B could also
work, but have more management issues involved
than D. C is not an option to connect DOSbased clients to a Windows NT server. See the
section titled "NetBEUI."
6. B. TCP/IP is the protocol that would have to be
used, because it is the only protocol that can be
used on the Internet. See the section titled
"TCP/IP—The Internet Protocols."
12. B. UNC names are always based on the
following:
\\Computername\Sharename\filename
See the section titled "NetBIOS Names."
13. C. To see the list of shared resources on a specific
computer from the command prompt, type:
net view\\Computername
See the section titled "NetBIOS Names."
Suggested Readings and Resources
1. Heywood, Drew. Networking with Microsoft
TCP/IP, 2nd Edition. New Riders, 1997.
3. Siyan, Karanjit S. Windows NT Server 4
Professional Reference. New Riders, 1996.
2. Dulaney, Emmett. MCSE Training Guide:
TCP/IP. New Riders, 1998.
4. Heywood, Drew. Inside Windows NT Server 4,
Administrators Resource Edition. New Riders,
1997.
13.919-x CH08.i
8/28/98 1:22 PM
Page 293
OBJECTIVES
Chapter 8 targets the following objective in the
Planning section of the Networking Essentials exam:
List the characteristics, requirements, and appropriate situations for WAN connection services.
WAN connection services include: X.25, ISDN,
Frame Relay, and ATM
. This is an important topic because it not only
applies the theory learned in Chapter 2, “Networking Standards,” but when you face some form
of WAN connectivity need or problem, one or
more of these options will more than likely be your
solution.
C H A P T E R
8
Connection Services
13.919-x CH08.i
8/28/98 1:22 PM
Page 294
OUTLINE
S T U DY S T R AT E G I E S
The Public Telephone Network
295
Leased Line Types
T1 and T3
Digital Data Service (DDS)
Switched 56
298
298
299
299
Packet Routing Services
299
ISDN and B-ISDN
301
X.25
302
Frame Relay
304
Asynchronous Transfer Mode (ATM)
305
Synchronous Optical Network (SONET)
307
Switched Multimegabit Digital Service
(SMDS)
307
Asymmetric Digital Subscriber Line
(ADSL)
307
Cable Modems
308
Chapter Summary
310
. The way to study for this material is to know the
characteristics of each of the various WAN technologies or services, including those crucial to
making decisions about which to implement.
These characteristics include such things as
speed, cost, and availability.
13.919-x CH08.i
8/28/98 1:22 PM
Page 295
C hapter 8
INTRODUCTION
Communication must occur between distant points, but few organizations can justify the costs required to construct a private wide area
network. Up to this point, all the chapters have focused primarily on
issues that relate to a local area network. As defined earlier in
Chapter 1, a LAN is characterized by high bandwidth and the fact
that the company controls and maintains all the connectivity devices
as well as the transmission media. This chapter discusses some of the
WAN connectivity issues that a company may wish to address.
Chapter 3, “Transmission Media,” discussed some of the possible
transmission media that a company could use to establish WAN
connectivity. These include spread-spectrum, infrared, and satellite
communications. Another possibility is fiber-optic cable, but often a
company does not have access to the property to lay down any form
of physical medium. All these options are expensive to undertake.
Fortunately, a variety of commercial options are available that enable
organizations to pay for only the level of service they require. These
commercial options take advantage of existing infrastructures supplied by the telephone companies, cable companies, and Internet
Service Providers (ISPs). This chapter discusses some wide area network (WAN) service options. You will also learn about dial-up versus dedicated service. This chapter describes some of the available
types of digital communication lines and examines some standards
for WAN connection services.
THE PUBLIC TELEPHONE NETWORK
A major issue with WAN connectivity is whether the level of service
you wish to employ exists at all points of communication. For example, if you are requiring a dedicated connection between Moscow,
Russia, and Santiago, Chile, there would have to be the same level
of service between both points of access. One WAN connectivity
service that exists almost worldwide is the public telephone system.
Although recently cable television companies have begun to provide
WAN connectivity service, almost all public carrier services to this
point have been offered by the telephone companies. Thus in this
CONNECTION SERVI C E S
295
13.919-x CH08.i
296
8/28/98 1:22 PM
C h a p te r 8
Page 296
C O NNE C TI O N S E RVICES
chapter, almost all the technologies discussed will address the public
telephone system.
Public telephone networks offer two general types of service:
á Leased dedicated services. The customer is granted exclusive
access.
á Dial-up services. The customer pays on a per-use basis.
Switched services operate the Public Switched Telephone Network
(PSTN), which you know as the telephone system. Voice-grade services have evolved to high levels of sophistication and can be adapted
to provide many data services with devices such as modems. Newer
switched options provide higher levels of service while retaining the
advantages of switched access.
With dial-up service, subscribers don’t have exclusive access to a particular data path. The PSTN maintains large numbers of paths but
not nearly enough to service all customers simultaneously. The most
obvious illustration of this is when you try to use the telephone and
a recorded message says “all lines are busy.” When a customer
requests service, a dedicated path is switched into service to meet the
customer’s needs. When the customer hangs up, the path is dissolved, and the circuits are available for use by other customers. In
situations in which the customer doesn’t need full-time network
access, switched service is extremely cost-effective.
The cost for this service can be either a flat monthly fee, or a time
used/distance (traveled fee long distance charges). These costs vary
considerably depending on where in the world you are located.
The workings of the public telephone system are seen in Figure 8.1.
Essentially any telephone or modem that uses a telephone jack connects into the wall. From the wall, lines (in most cases UTP lines)
run to a demarc unit within the building. A demarc unit is the connection point between all the telephone lines in a building and the
local loop line that leaves the building.
The local loop lines are usually higher-grade UTP or fiber-optic
wires that are owned and maintained by the telephone company.
These local loop lines connect the building with the Central Office
(CO). The central office is the local telephone station point that
switches calls from one local loop line to another, to a trunk line, or
13.919-x CH08.i
8/28/98 1:22 PM
Page 297
C hapter 8
CONNECTION SERVI C E S
297
FIGURE 8.1
The elements that comprise a Public Switched
Telephone Network.
Demarc
Local Loop
Subscriber
Premise Wiring
Central Office
(CO)
Long Distance
Carrier
High-Capacity
Trunk Line
Central Office
to a long distance carrier. Trunk lines connect COs together. Local
loop lines are higher in bandwidth than the wires running within a
building. Trunk lines are even higher in bandwidth than local loop
lines.
It is due to all this switching that the telephone system is referred to
as the Public Switched Telephone Network. You must remember,
however, that a telephone line connection is a circuit switching (refer
to Chapter 2) technology. That is, a constant path is established
between the two devices on the network.
When you use a modem over the telephone system, the modem
converts the computer’s digital signal to an analog signal. This analog signal is then passed over an established pathway to the device
Non-Local
Central Office
13.919-x CH08.i
298
8/28/98 1:22 PM
C h a p te r 8
Page 298
C O NNE C TI O N S E RVICES
on the other end. You are charged for data transfer, if dialing long
distance, whether or not data throughput is happening. This is
because the telephone system is renting you constant line access.
Leased Line Types
When customers require full-time access to a communication path, a
dedicated, leased line serves as one option. Several levels of digital
lines are available. Digital lines are superior to analog lines because
they tend to be more error-free. Following are some examples of digital line services:
á T1 and T3
á Digital data service
á Switched 56
T1 and T3
A very popular digital line is the T1 leased line. This leased line
provides point-to-point connections and transmits a total of 24
channels across two wire pairs—one pair for sending and one for
receiving—for a transmission rate of 1.544Mbps. A T1 is known as
an E1 line in Europe.
Very few private networks require the capacity of a T1 line. The
channels of a T1 line are often leased out on a fractional basis. Each
T1 channel can transmit up to 64Kbps of data. All 24 channels at
once equal 1.544Mbps. Fractional T1s are often guaranteed at
56Kbps, with 8Kbps being set aside for management purposes.
T3 (E3 in Europe) is similar to T1, but T3 has an even higher
capacity. In fact, a T3 line can transmit at up to 45Mbps. This is
because a T3 line is made up of 672 64Kbps channels.
A single-channel service on a T1 is called DS-0. DS-1 service is a full
T1 line. DS-1C is two T1 lines, DS-2 is four T1 lines, and DS-3 is a
full T3 line (equivalent to 28 T1s). A level of service called T4 is
equal to 168 T1 lines.
13.919-x CH08.i
8/28/98 1:22 PM
Page 299
C hapter 8
Digital Data Service (DDS)
Digital Data Service (DDS) is a very basic form of digital service.
DDS transmits from point to point at 2.4, 4.8, 9.6, or 56Kbps. In
its most basic form, DDS provides a dedicated line.
Switched 56
A special service related to DDS, Switched 56 offers a dial-up version
of the 56Kbps DDS. With Switched 56, users can dial other
Switched 56 sites and pay for only the connect time.
PACKET ROUTING SERVICES
Many organizations must communicate among several points.
Leasing a line between each pair of points can prove too costly.
Many services now are available that route packets between different
sites. Some of the packet-routing services discussed in this chapter
are as follows:
á ISDN
á X.25
á Frame Relay
á ATM
á SONET
á SMDS
á ADSL
á Cable Modem
Each of these services has characteristics that suit it to particular
uses, and all these services are available on a leased basis from service
providers, yet these services are not available in all locations. An
organization that must communicate among many sites simply pays
to connect each site to the service, and the service assumes the
responsibility of routing packets. The expense of operating the network is then shared among all network subscribers. Because the
exact switching process is concealed from the subscriber, these networks frequently are depicted as a communication cloud, as shown
in Figure 8.2.
CONNECTION SERVI C E S
299
13.919-x CH08.i
300
8/28/98 1:22 PM
C h a p te r 8
Page 300
C O NNE C TI O N S E RVICES
FIGURE 8.2
How a public network service is often represented.
LAN
NOTE
Public
Data Network
Service
Multiplexing Many digital transmission methods use a technique called
multiplexing. Multiplexing, described in
Chapter 3 enables broadband media
to support multiple data channels.
The data rates of public network services can be compared to common LAN services such as Ethernet (10–100Mbps) and Token Ring
(4–16Mbps) to give you an idea of how the public services’ speed
affects performance of the network’s communications.
Before going into detail on the following WAN connectivity services,
a recap of packet routing concepts is warranted. In Chapter 2, you
were introduced to packet switching and other routing-related techniques used to send data over WAN links. Packet-switching networks often use virtual circuits to route data from the source to the
destination. A virtual circuit is a specific path through the network—
a chain of communication links leading from the source to the destination (as opposed to a scheme in which each packet finds its own
path). Virtual circuits enable the network to provide better errorchecking and flow control.
The two main forms of virtual circuits are the following:
á A switched virtual circuit (SVC) is created for a specific com-
munication session and then disappears after the session. The
13.919-x CH08.i
8/28/98 1:22 PM
Page 301
C hapter 8
CONNECTION SERVI C E S
301
next time the computers communicate, a different virtual circuit might be used.
á A permanent virtual circuit (PVC) is a permanent route
through the network that is always available to the customer.
With a PVC, charges are still billed on a per-use basis.
ISDN and B-ISDN
Integrated Services Digital Network (ISDN) is a group of ITU
(CCITT) standards designed to provide voice, video, and data transmission services on digital telephone networks. ISDN uses multiplexing to support multiple channels on high-bandwidth circuits.
The relationship of the ISDN protocols to the OSI reference model
is shown in Figure 8.3.
The original idea behind ISDN was to enable existing phone lines
to carry digital communications, and was at one time touted as a
replacement to traditional analog lines. Thus, ISDN is more like
traditional telephone service than some of the other WAN services
discussed further on in this chapter. ISDN is intended as a dial-up
service and not as a permanent, 24-hour connection.
ISDN separates the bandwidth into channels (see the following “Indepth” for more information). Based upon how these channels are
used, ISDN can be separated into two classes of service.
á Basic Rate (BRI). Basic Rate ISDN uses three channels. Two
channels (called B channels) carry the digital data at 64Kbps. A
third channel (called the D channel ) provides link and signaling information at 16Kbps. Basic Rate ISDN thus is referred to
as 2B+D. A single PC transmitting through ISDN can use
both B channels simultaneously, providing a maximum data
rate of 128Kbps (or higher with compression).
á Primary Rate (PRI). Primary Rate supports 23 64Kbps B chan-
nels and one 64Kbps D channel. The D channel is used for
signaling and management, whereas the B channels provide
the data throughput.
With a BRI line, if the line is currently being used for voice, only
one B channel is available for data. This effectively reduces the
throughput of the BRI down to 64Kbps.
Application
Presentation
Session
Transport
ISDN
Network
Data Link
LAPD
Physical
FIGURE 8.3
The relationship between ISDN and the OSI reference model.
13.919-x CH08.i
302
8/28/98 1:22 PM
C h a p te r 8
Page 302
C O NNE C TI O N S E RVICES
I S DN CHA NNE L TY P ES
A variety of ISDN channel types are defined. These channel types,
often called bit pipes, provide different types and levels of service.
The following list details the various channels:
á A channels. Provide 4KHz analog telephone service.
á B channels. Support 64Kbps digital data.
á C channels. Support 8 or 16Kbps digital data, generally for
out-of-band signaling.
á D channels. Support 16 or 64Kbps digital data, also for outof-band signaling. D channels support the following subchannels:
• p subchannels support low-bandwidth packet data.
• s subchannels are used for signaling (such as call
setup).
• t subchannels support telemetry data (such as utility
meters).
á E channels. Provide 64Kbps service used for internal ISDN
signaling.
á H channels. Provide 384, 1,536, or 1,920Kbps digital service.
ISDN functions as the data-transmission service only. The LAPD
protocol, which operates on the D channel, provides the acknowledged, connectionless, full-duplex service, as well as the physical
addressing service.
Broadband ISDN (B-ISDN) is a refinement of ISDN that is defined
to support higher-bandwidth applications, such as video, imaging,
and multimedia. Physical layer support for B-ISDN is provided by
Asynchronous Transfer Mode (ATM) and the Synchronous Optical
Network (SONET), both discussed later in this chapter. Typical BISDN data rates are 51Mbps, 155Mbps, and 622Mbps over fiberoptic media.
X.25
X.25 is a packet-switching network standard developed by the
International Telegraph and Telephone Consultative Committee
13.919-x CH08.i
8/28/98 1:22 PM
Page 303
C hapter 8
(CCITT, abbreviated from the French), which has been renamed the
International Telecommunications Union (ITU). The standard,
referred to as Recommendation X.25, was introduced in 1974 and is
now implemented most commonly in WANs.
As shown in Figure 8.4, X.25 is one level of a three-level stack that
spans the Network, Data Link, and Physical layers. The middle
layer, Link Access Procedures-Balanced (LAPB), is a bit-oriented, fullduplex, synchronous Data Link layer LLC protocol. Physical layer
connectivity is provided by a variety of standards, including X.21,
X.21bis, and V.32.
X.25 packet-switching networks provide the options of permanent
or switched virtual circuits. Although a datagram (unreliable) protocol was supported until 1984, X.25 is now required to provide reliable service and end-to-end flow control. Because each device on a
network can operate more than one virtual circuit, X.25 must provide error and flow control for each virtual circuit.
At the time X.25 was developed, this flow control and error checking was essential because X.25 was developed around relatively unreliable telephone line communications. The drawback is that error
checking and flow control slow down X.25. Generally, X.25 networks are implemented with line speeds of up to 64Kbps, although
actual throughput seems slower due to the error correction controls
in place. These speeds are suitable for the file transfer and terminal
activity that comprised the bulk of network traffic when X.25 was
defined, most of this traffic being terminal connections to mainframes. Such speeds, however, are inadequate to provide LAN-speed
services, which typically require speeds of 1Mbps or better. X.25
networks, therefore, are poor choices for providing LAN application
services in a WAN environment. One advantage of X.25, however, is
that it is an established standard that is used internationally. This, as
well as lack of other services throughout the world, means that X.25
is more of a connection service to Africa, South America, and Asia,
where lack of other services prevail.
Figure 8.5 shows a typical X.25 configuration. In X.25 parlance, a
computer or terminal is called data terminal equipment (DTE ). A
DTE can also be a gateway providing access to a local network.
Data communications equipment (DCE) provides access to the
packet-switched network (PSN ). A PSE is a packet-switching exchange,
also called a switch or switching node.
CONNECTION SERVI C E S
Application
Presentation
Session
Transport
Network
X.25
Data Link
LAPB
Physical
X.21, etc.
FIGURE 8.4
The relationship between X.25 and the OSI
reference model.
303
13.919-x CH08.i
304
8/28/98 1:23 PM
C h a p te r 8
Page 304
C O NNE C TI O N S E RVICES
FIGURE 8.5
PDN (Public Digital Network)
A sample X.25 network.
DCE
DTE
PAD
DCE
PSE
DCE
PSE
DCE
PAD
DTE
The X.25 protocol oversees the communication between the DTE
and the DCE. A device called a packet assembler/disassembler (PAD)
translates asynchronous input from the DTE into packets suitable
for the PDN.
Frame Relay
Frame relay was designed to support the Broadband Integrated
Services Digital Network (B-ISDN), which was discussed in the preceding section. The specifications for frame relay address some of the
limitations of X.25. As with X.25, frame relay is a packet-switching
network service, but frame relay was designed around newer, faster
fiber-optic networks.
Application
Presentation
Unlike X.25, frame relay assumes a more reliable network. This
enables frame relay to eliminate much of the X.25 overhead required
to provide reliable service on less reliable networks. Frame relay relies
on higher-level protocol layers to provide flow and error control.
Session
Transport
Network
Data Link
Physical
Frame
Relay
FIGURE 8.6
The relationship between Frame Relay and the
OSI reference model.
Frame relay is typically implemented as a public data network and,
therefore, is regarded as a WAN protocol. The relationship of Frame
Relay to the OSI model is shown in Figure 8.6. Notice that the
scope of Frame Relay is limited to the Physical and Data Link layers.
Frame relay provides permanent virtual circuits that supply permanent virtual pathways for WAN connections. Frame relay services
13.919-x CH08.i
8/28/98 1:23 PM
Page 305
C hapter 8
CONNECTION SERVI C E S
305
are typically implemented at line speeds from 56Kbps up to
1.544Mbps (T1).
Customers typically purchase access to a specific amount of bandwidth on a frame relay service. This bandwidth is called the committed information rate (CIR), a data rate for which the customer is
guaranteed access, and is available in increments of 64Kbps.
Customers might be permitted to access higher data rates on a
pay-per-use, temporary basis. This arrangement enables customers
to tailor their network access costs based on their bandwidth
requirements.
To use frame relay, you must have special, frame-relay-compatible
connectivity devices (such as frame-relay-compatible routers and
bridges).
Asynchronous Transfer Mode (ATM)
Asynchronous Transfer Mode (ATM) is a high-bandwidth switching
technology developed by the ITU Telecommunications Standards
Sector (ITU-TSS). An organization called the ATM Forum is
responsible for defining ATM implementation characteristics. ATM
can be layered on other Physical layer technologies, such as Fiber
Distributed Data Interface (FDDI) and SONET. The relationships of
these protocols to the OSI model are shown in Figure 8.7.
Several characteristics distinguish ATM from other switching technologies. ATM is based on fixed-length, 53-byte cells, whereas other
technologies employ frames that vary in length to accommodate different amounts of data. Because ATM cells are uniform in length,
switching mechanisms can operate with a high level of efficiency.
This high efficiency results in high data transfer rates. Some ATM
systems can operate at an incredible rate of 622Mbps; a typical
working speed for an ATM is around 155Mbps.
The unit of transmission for ATM is called a cell. All cells are 53
bytes long and consist of a 5-byte header and 48 bytes of data.
The 48-byte data size was selected by the standards committee as
a compromise to suit both audio- and data-transmission needs.
Audio information, for instance, must be delivered with little
latency (delay) to maintain a smooth flow of sound. Audio engineers
therefore preferred a small cell so that cells would be more readily
Application
Presentation
Session
Transport
Network
Data Link
Physical
ATM
SONET/SDH, FDDI, etc.
FIGURE 8.7
The relationship of ATM to the OSI reference
model.
13.919-x CH08.i
306
8/28/98 1:23 PM
C h a p te r 8
Page 306
C O NNE C TI O N S E RVICES
available when needed. For data, however, large cells reduce the overhead required to deliver a byte of information.
Asynchronous delivery is another distinguishing feature of ATM.
Asynchronous refers to the characteristic of ATM in which transmission time slots don’t occur periodically but are granted at irregular
intervals. ATM uses a technique called label multiplexing, which allocates time slots on demand. Traffic that is time-critical, such as voice
or video, can be given priority over data traffic that can be delayed
slightly with no ill effect. Channels are identified by cell labels, not
by specific time slots. A high-priority transmission need not be held
until its next time slot allocation. Instead, it might be required to
wait only until the current 53-byte cell has been transmitted.
TI ME -DI V I S I ON MULTI P LE X I NG
Other multichannel technologies utilize time-division techniques to
allocate bandwidth to channels. A T1 (1.544Mbps) line, for example, might be time-division multiplexed to provide 24 voice channels. With this technique, each channel is assigned a specific time
slot in the transmission schedule. The disadvantage of this technique is that an idle channel doesn’t yield its bandwidth for the creation of other channels.
Devices communicate on ATM networks by establishing a virtual
path, which is identified by a virtual path identifier (VPI). Within
this virtual path, virtual circuits can be established, which are in turn
associated with virtual circuit identifiers (VCIs). The VPI and VCI
together make up a 3-byte field included in the cell header.
ATM is relatively new technology, and only a few suppliers provide
the equipment necessary to support it. (ATM networks must use
ATM-compatible switches, routers, and other connectivity devices.)
Other networks, such as a routed ethernet, require a 6-byte physical
address as well as a network address to uniquely identify each device
on an internetwork. An ATM can switch cells with 3-byte identifiers
because VPIs and VCIs apply to only a given device-to-device link.
Each ATM switch can assign different VPIs and VCIs for each link,
and up to 16 million circuits can be configured for any given deviceto-device link.
13.919-x CH08.i
8/28/98 1:23 PM
Page 307
C hapter 8
Although ATM was developed primarily as a WAN technology, it
has many characteristics of value for high-performance LANs. An
interesting advantage of ATM is that ATM makes it possible to use
the same technology for both LANs and WANs. Some disadvantages, however, include the cost, the limited availability of the equipment, and the present lack of expertise regarding ATM due to its
relatively recent arrival.
Synchronous Optical Network (SONET)
Bell Communications Research developed SONET, which has been
accepted as an ANSI standard. As the “optical” in the name implies,
SONET is a standard for communication over fiber-optic networks.
Data rates for SONET are organized in a hierarchy based on the
Optical Carrier (OC) speed and the corresponding Synchronous
Transport Signals (STS) employed. The basic OC and STS data rate
is 51.84Mbps, but higher data rates are provided in multiples of the
basic rate. Thus OC-48 is 48´51.84Mbps or 2488.32 Mbps.
Switched Multimegabit Digital Service
(SMDS)
Developed by Bell Communications Research in 1991, SMDS technology is related to ATM in that it transports data in 53-byte cells.
SMDS (see Figure 8.8) is a connectionless Data Link layer service
that supports cell switching at data rates of 1.544 to 45Mbps. IEEE
802.6 (DQDB metropolitan area network) is the primary Physical
layer standard employed with SMDS, although other Physical layer
standards are supported.
Asymmetric Digital Subscriber Line
(ADSL)
One new type of broadband WAN connectivity being tested by the
telephone companies is ADSL. Available since only 1997, ADSL is a
Physical layer standard of sending data across existing telephone
wires. By using a special ADSL modem, users can receive data at
CONNECTION SERVI C E S
307
13.919-x CH08.i
308
8/28/98 1:23 PM
C h a p te r 8
Page 308
C O NNE C TI O N S E RVICES
FIGURE 8.8
The relationship of SMDS to the OSI Reference
model.
Application
Presentation
Session
Transport
Network
SMDS
Data Link
Physical
Sonet
rates over 8Mbps, and send data at rates of up to 640Kbps. This is
accomplished through the use of Frequency Division Multiplexing
across the existing telephone lines. There is supposed to be support
for ATM and IP protocols.
Cable Modems
Another new area of expansion into WAN connectivity services is
the advent of the cable modem. This device enables networks to
interconnect through existing cable TV lines. Some areas that offer
this service have a full-duplex version that is capable of transmitting
data at rates of 4–10Mbps. Other areas have cable standards in place
that enable the coaxial TV cable to only receive data, relying upon
an analog dial-up connection to be used to send data. This is definitely another area of technology that should be watched closely over
the next few years.
13.919-x CH08.i
8/28/98 1:23 PM
Page 309
C hapter 8
C A S E S T U DY : M A K I N G D E C I S I O N S
SERVICES OPTIONS
ON
CONNECTION SERVI C E S
309
CONNECTIVITY
ESSENCE OF THE CASE
A N A LY S I S
The essence of the case is as follows:
The first two areas of analysis should be to
decide whether Berg Industries needs point-topoint support or some form of multipoint support
and what type of service is available in each of
the three areas.
• Berg Industries has three locations.
• They wish to have bandwidth support for
large data transfers.
• They wish to have WAN bandwidth support for email.
SCENARIO
Berg Industries (BI), a major gold explorer, has
called you in to consult on their network connectivity. BI currently has three locations: one in
Vancouver, Canada, another in San Francisco,
and a third in Lima, Peru.
BI is heavily involved in surveying and mapping
gold reserves found in Canada, the United
States, and Peru. Each of these offices has its
own LAN. Each local LAN has two main servers.
One server houses the geological exploration
data while the other server is the email server.
BI would like all geological data to be stored on
the Vancouver server, for archival purposes. This
data usually amounts to over a hundred
megabytes of data every week.
BI would also like to have connectivity with the
email systems. Email does not typically use a lot
of bandwidth. You did, however, notice that all the
local desktop PCs were fully multimedia
equipped, and learned after talking to a few of
the staff that small avi files are often sent
around by staff members.
Your job is to present to BI the various commercial connectivity options that it could pursue to
meet its needs.
Because BI has only three locations, and theoretically only Vancouver needs to connect to San
Francisco and Lima, point-to-point connections
are an option. This would mean that all email
would be routed through Vancouver, so if
Vancouver’s link goes down, email cannot go
between Lima and San Francisco. In any case, a
multipoint or a point-to-point system can both
perform the function. Often when a company has
multiple locations, or multiple nodes per location
that must be connected, a multipoint option wins
out over a point-to-point connection option.
The second area of concern is what services are
available in what areas. Further investigation
reveals that between San Francisco and
Vancouver, PSTN, ISDN, X.25, and T1 are all available. Between Lima and Vancouver, as well as
Lima and San Francisco, only the PSTN and X.25
systems are available.
The next area of concern is to break the data
transmissions down into major types or categories. The two main types of data transfer BI
engages in are those of email and geological
data transfers.
As mentioned earlier, the geological data files are
quite large. In fact, they are often over 100
megabytes or greater. This type of daily data
transfer definitely requires a lot of bandwidth.
Given the options, a T1 line would work the best,
continues
13.919-x CH08.i
310
8/28/98 1:23 PM
C h a p te r 8
Page 310
C O NNE C TI O N S E RVICES
C A S E S T U DY : M A K I N G D E C I S I O N S
SERVICES OPTIONS
ON
CONNECTIVITY
continued
distance rates apply. A company that communicates a lot may find that the speed of
the X.25 service in delivering the emails
justifies any small additional expense that
X.25 may have over the PSTN.
given that it can transfer at up to 1.544Mbps.
The problem is that no T1 service runs down to
Lima.
As for the email, the X.25 and the PSTN would
suffice. Both of these services are available. The
main difference between the two options is that
the PSTN is based on a dialup connection. Thus
email would probably be delayed, because most
email systems wait to dial up every few hours, or
dial up only after a certain number of messages
has been queued. With X.25, you have a constant connection, so email gets delivered almost
as soon as it is sent.
Before any final answer can be determined, the
following three issues would need to be discussed with Berg Industries management:
• How quickly the geological data needs to be
sent to Vancouver. A T1 line is rather expensive. In fact, it would probably be cheaper
to dump the data onto a CD or tape backup, and courier a copy to Vancouver every
week.
• The cost of the PSTN versus the X.25 service. A PSTN is on timed intervals, and long
• How quickly an X.25 link could be established in Peru. In North America, people
often expect and receive very fast service.
In many third world countries, it can take
months (even years in some cases) to get
services hooked up. Thus the time it would
take to get an X.25 link established should
also be considered.
This case study, like so many of the others, has
shown that an obvious solution is not always that
easy to find. Often you need to gather more information than the situation may lend itself to or
you come to realize that tradeoffs must be made
between costs and efficiency. You present the
client with the possibilities and then help them
make the decision that works best for them,
even if it is not the “perfect” solution. The perfect solution can be beyond most corporate budgets or the available technology.
CHAPTER SUMMARY
KEY TERMS
• PSTN
This chapter examined some basic WAN connectivity concepts, such
as dial-up and dedicated service lines. You learned about some types
of digital lines, such as the following:
• leased lines
á T1 and T3
• T1
á DDS
• T3
á Switched 56
13.919-x CH08.i
8/28/98 1:23 PM
Page 311
C hapter 8
CONNECTION SERVI C E S
311
CHAPTER SUMMARY
This chapter also described the characteristics of some important
WAN connectivity service standards and appropriate situations for
utilizing them. The services included the following:
• Digital Data Service (DDS)
• Switched 56
á ISDN
• switched virtual circuit (SVC)
á X.25
• Permanent Virtual Circuit (PVC)
á Frame Relay
á ATM
• Integrated Services Digital
Network (ISDN)
á SMDS
• Basic Rate ISDN (BRI)
á SONET
• Primary Rate ISDN (PRI)
The services also included the following two new technologies
emerging on the WAN horizon:
á ADSL
• X.25
• Frame Relay
á Cable Modems
For more information on packet switching and virtual circuits, refer
to Chapter 2.
The WAN connection services are summarized in Table 8.1.
• Asynchronous Transfer Mode
(ATM)
• Synchronous Optical Network
(SONET)
• Switched Multimegabit Digital
Service (SMDS)
TABLE 8.1
WAN C O N N E C T I O N C O M PA R I S O N
Service
• Broadband ISDN (B-ISDN)
Speed
Connection Type
Connection Format
PSTN
up to 56Kbps
Dial-up
Point to Point
T1
1.544Mbps
Permanent
Point to Point
T3
45Mbps
Permanent
Point to Point
DDS
up to 56Kbps
Permanent
Point to Point
Switched 56
up to 56Kbps
Permanent
Point to Point
ISDN
128Kbps
Dial-up
Point to Point
B-ISDN
51–622Mbps
Dial-up
Point to Point
X.25
56Kbps
Permanent
Multipoint
Frame Relay
1.544Mbps
Permanent
Multipoint
ATM
up to 622Mbps
Permanent
Multipoint
SMDS
45Mbps
Permanent
Multipoint
SONET
2488Mbps
Permanent
Multipoint
ADSL
8Mbps
Permanent
Multipoint
Cable Modem
10Mbps
Permanent
Multipoint
• Asymmetric Digital Subscriber
Line (ADSL)
• cable modem
13.919-x CH08.i
312
8/28/98 1:23 PM
C h a p te r 8
Page 312
C O NNE C TI O N S E RVICES
A P P LY Y O U R L E A R N I N G
Exercises
8.1
Accessing an X.25 Network Through
Windows NT Dial-Up Networking
Objective: Learn how to configure Windows NT DialUp Networking to connect to an X.25 network
provider.
Estimated time: 15 minutes
Windows NT Remote Access Service (RAS) is usually
used for modem connections to remote PCs, but you
can also use RAS to access an X.25 packet-switching
network. RAS supports Packet Assembler/Disassembler
(PAD) devices and X.25 smart cards. Alternatively, you
can use Windows NT’s Dial-Up Networking to connect to a commercial X.25 provider. This exercise
assumes that any modems or RAS are not already
installed. The purpose of this exercise is to simply show
how you would go about selecting an X.25 PAD device
in RAS.
1. Click the Start menu and choose Settings/
Control Panel. Double-click the Windows NT
Control Panel Network application.
2. Choose the Network application’s Services tab.
Click on the Add button. Choose Remote Access
Service from the Network Services list and click
the OK button. You are prompted for the location of the installation files. Type in the location
and click on the Continue button.
3. A dialog box appears telling you that there are no
RAS-compatible devices to add, and asking
whether you wish to enable the Modem installer
program to add a modem. Click on the No button.
4. The Add RAS Device dialog box appears. Click
on the Install X.25 Pad button.
5. The Install X25 PAD dialog box appears. Here
you need to select the port to which the PAD is
to be connected, and the X.25 PAD name. This
PAD name must correspond to the PAD supplied
to you by your service provider. Select a PAD
device and click on OK.
6. Click on OK again, and then in the Remote
Access Setup click on Cancel. This stops the RAS
install. You do not wish to actually install the
PAD device because you do not have one connected to your machine. You are prompted that
your changes will not be saved. Click on the Yes
button.
7. Click on the Cancel button to close the Network
applet.
Review Questions
1. What are the main differences between dial-up
and leased lines?
2. X.25 has one main benefit over other WAN connectivity options. What is it?
3. ATM is fast. What are its drawbacks?
Exam Questions
1. To what does the D channel refer in ISDN?
A. Data rate
B. Degradation signaling
C. 16Kbps control channel
D. 144Kbps combination channel
13.919-x CH08.i
8/28/98 1:23 PM
Page 313
C hapter 8
CONNECTION SERVI C E S
313
A P P LY Y O U R L E A R N I N G
2. The DS-0 service level provides a transmission
rate of what speed?
A. 64Kbps
B. 128Kbps
A. X.25
B. ISDN
C. ATM
D. Frame relay
C. 1.544Mbps
D. 45Mbps
7. What is sometimes called 2B+D?
A. Primary rate ISDN
3. A T3 line provides a transmission rate of what
speed?
A. 64Kbps
B. 128Kbps
C. 1.544Mbps
D. 45Mbps
4. What is an SVC?
A. It is a permanent path. Charges are billed on
a monthly basis.
B. It is a permanent path. Charges are billed on
a per-use basis.
C. It is a temporary path created for a specific
communication session.
D. It is none of the above.
B. Basic rate X.25
C. Primary rate frame relay
D. Basic rate ISDN
8. What is a typical working speed for ATM?
A. 1.544Mbps
B. 45Mbps
C. 155Mbps
D. 622Mbps
9. What is the size and name of the byte-size blocks
in which ATM divides data?
A. 53, packets
B. 53, cells
C. 56, frames
5. How does X.25 compare to Frame Relay?
D. 128, cells
A. Faster than
B. Slower than
C. About the same speed as
D. Nearly identical to
10. A modem uses what type of signaling?
A. Digital
B. Analog
C. Dedicated
6. What was designed to provide digital communications over existing phone lines?
D. None of the above
13.919-x CH08.i
314
8/28/98 1:23 PM
C h a p te r 8
Page 314
C O NNE C TI O N S E RVICES
A P P LY Y O U R L E A R N I N G
11. Which three of the following are digital line
options?
A. Switched 16
B. T1
C. DDS
D. Switched 56
12. What service uses a PAD?
A. X.25
Suggested Solution: You utilize an X.25 connection service.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objective.
D. This solution does not satisfy the primary
objective.
B. Frame relay
C. ATM
D. ISDN
13. What service transfers data in fixed-length units
called cells?
A. X.25
B. Frame relay
C. ATM
D. ISDN
14. The corporate network needs to expand and connect over a WAN to many different international
locations. You need to come up with a solution
that will enable the corporation to communicate.
Primary objective: You need constant connectivity.
Secondary objective: Data transfer speeds of
56Mbps are desired.
Secondary objective: You need to utilize a standard that is available worldwide.
Answers to Review Questions
1. Dial-up lines are used and accessed only when the
subscriber initializes them. They tend to be
cheaper, because a circuit is established only when
the subscriber initializes the call. When a subscriber is not using the service, the lines are available for other subscribers. PSTN and ISDN lines
are examples of this.
A leased line is a constant connection that is dedicated to a subscriber. You are paying for the 24hour-a-day service, as well as a possible data
throughput usage. A T1 line is an example of
this.
See the sections titled “ISDN and B-ISDN” and
“The Public Telephone Network.”
2. The main advantage of X.25 is that it is an
accepted standard utilized by almost all telephone
companies worldwide. Thus its global availability
is its main benefit. See the section titled “X.25.”
3. ATM is very fast—up to 622Mbps. Its drawbacks
are its price and limited availability. See the section titled “ATM.”
13.919-x CH08.i
8/28/98 1:23 PM
Page 315
C hapter 8
CONNECTION SERVI C E S
315
A P P LY Y O U R L E A R N I N G
Answers to Exam Questions
1. C. The D channel on an ISDN line is used for
controlling and signaling the transmission. See
the section titled “ISDN.”
2. A. A DS-0 is a single channel on a T1 line. Each
channel on a T1 line has a speed of 64Kbps. See
the section titled “T1 and T3.”
3. D. A single channel in a T1 and T3 line provides
64Kbps. A T1’s total capacity is 1.544Mbps. See
the section titled “T1 and T3.”
8. C. ATM typically operates at 155Mbps, but can
reach speeds of up to 622Mbps. See the section
titled “ATM.”
9. B. ATM’s speed is due to the uniform size of its
data cells. Each cell is 53 bytes in length. See the
section titled “ATM.”
10. B. A modem uses analog signals. Modem stands
for MODulator/DEModulator. See the section
titled “Digital and Analog Signaling” in Chapter
5, “Network Adapter Cards.”
4. C. SVC stands for Switched Virtual Circuit. See
the section titled “X.25.”
11. B, C, D. All options except A are digital line
options. There is nothing called “Switched 16.”
See the section titled “Leased Line Types.”
5. B. X.25 is slower than frame relay. X.25 has
speeds of up to 56Kbps, whereas frame relay’s
speeds approach 1.544Mbps. See the sections
titled “X.25” and “Frame Relay.”
12. A. A PAD is a Packet Assembly Device. This is
the connecting unit for an X.25 network. Frame
Relay, ATM, and ISDN do not use a PAD. See
the section titled “X.25.”
6. B. ISDN was designed to provide digital communications over existing phone lines. X.25 is a
packet-switching standard. ATM and frame relay
are not designed to go over existing phone lines.
See the section titled “ISDN.”
13. C. ATM uses 53-byte packets of data called cells.
See the section titled “Asynchronous Transfer
Mode (ATM).”
7. D. 2B+D is BRI or Basic Rate ISDN. See the
section titled “ISDN.”
14. B. X.25 allows for constant connectivity and is
also available worldwide. The first secondary
objective cannot be met, however, because X.25
supports only up to 56Kbps, not 56Mbps. See
the section titled “X.25.”
Suggested Readings and Resources
1. Black, Ulysses. Computer Networks: Protocols,
Standards and Interfaces—The Professionals
Guide. Prentice Hall, 1993.
2. Horak, Ray, and Mark Miller. Communication
Systems and Networks: Voice, Data and
Broadband Technologies. IDG, 1993.
13.919-x CH08.i
8/28/98 1:23 PM
Page 316
14.919-x CH09.i
8/28/98 1:38 PM
Page 317
OBJECTIVES
Chapter 9 targets the following objective in the
Implementation section of the Networking Essentials
exam:
Choose a disaster recovery plan for various situations.
. This exam topic deals with proactive measures that
you can take to prevent lost data and server downtime.
C H A P T E R
9
Disaster Recovery
14.919-x CH09.i
8/28/98 1:38 PM
Page 318
OUTLINE
Protecting Data
S T U DY S T R AT E G I E S
319
Backup
319
Uninterruptible Power Supply
322
. When reading this chapter, pay particular attention to what type of options are available to protect data. Be especially aware of those options
that have to do with hard drive failures.
Recovering from System Failure
324
. Also be aware of the differences between the
different RAID levels of fault tolerance.
Implementing a Fault-Tolerant Design
Using RAID
Choosing a RAID Level
Disk Duplexing
324
324
328
329
Other Fault-Tolerance Mechanisms
331
Chapter Summary
335
14.919-x CH09.i
8/28/98 1:38 PM
Page 319
C hapter 9
INTRODUCTION
One of the major issues that a network administrator must address
is the possibility of system failure and associated downtime. The
administrator must handle two major issues to guard against the
danger of a failed server:
á Protecting data
á Reducing downtime
This chapter discusses both issues and examines how the use of
fault-tolerant disk configurations and a backup strategy can help
reduce the danger of lost time and data.
PROTECTING DATA
Natural disasters, equipment failures, power surges, and deliberate
vandalism can cause the catastrophic loss of precious network data.
Protecting the data is a primary responsibility of the network administrator. Microsoft highlights these important strategies for preventing data loss:
á Backup
á Uninterruptible Power Supply (UPS)
Both strategies are discussed in the following sections.
Backup
A backup schedule is an essential part of any data-protection strategy. You should design a backup system that is right for your situation and the data on your network.
A number of different strategies can be used in backing up files.
One way is simply to copy a file to another drive. Operating systems, however, typically have special backup commands that help
you with some of the bookkeeping required for maintaining a systematic backup schedule. Most backup commands mark the file
with the date and time of the backup so that you (and the backup
DISASTER RECOV E RY
319
14.919-x CH09.i
320
8/28/98 1:38 PM
C h a p te r 9
Page 320
D I S A S TE R R E C OV ERY
utility) can know when a copy of the file was last saved. This is the
purpose of the FAT file system’s Archive attribute. To determine
whether this attribute exists, check the properties of any file on a
FAT partition. If the Archive attribute is enabled, the file has
changed since the last time a backup was done. In this chapter, you
will see that some backup techniques reset this attribute, whereas
others do not.
Although backups can be accomplished by saving files to a different
drive, they typically are performed with some form of tape drive.
Commonly called DAT drives, these devices are capable of storing
many gigabytes of information quickly and economically. Moreover,
the tapes are small and portable and cheaper on a per-megabyte basis
than a hard drive. Another important step in your backup plan,
therefore, is deciding where to store these backup tapes. Many companies choose to make two copies of each backup tape and store one
of the copies off-site, thereby guarding against a catastrophic event
such as a fire.
In addition to two types of copy commands, Microsoft identifies the
following backup types:
á Full backup. Backs up all specified files.
á Incremental backup. Backs up only those files that have
changed since the last full or incremental backup.
á Differential backup. Backs up the specified files if the files have
changed since the last backup. This type doesn’t mark the files
as having been backed up, however. (A differential backup is
somewhat like a copy command. Because the file is not
marked as having been backed up, a later differential or incremental backup backs up the file again.)
á Daily Copy. This is a Microsoft Windows NT NTBACKUP
utility specific command. This command backs up only those
files that were changed the day that this option was selected
when doing a Daily Copy backup and does not modify the
archive bit of the files being backed up. This is a useful option
if you wish to do a backup outside the regular backup schedule
and do not wish to alter or affect the normal backup routine.
á Copy. This is the other Microsoft Windows NT NTBACKUP
utility specific command. This command backs up all selected
files, but does not modify the archive bit of those files being
14.919-x CH09.i
8/28/98 1:38 PM
Page 321
C hapter 9
backed up. Again, this is a useful option if you wish to do a
backup outside the regular backup schedule and do not wish
to alter or affect the normal backup routine.
A typical backup plan includes some combination of these backup
types performed at regular intervals. One common practice is to perform an incremental or differential backup each day and a full backup every week. Full backups make the restoration process easier
because there is theoretically only one set of tapes from which to
restore; however, they also require a lengthy backup process each
night. This could mean that if the backup tape media is not large
enough, someone must physically change the tapes, or there simply
may not be enough time in the night to perform a full backup of all
the data. Companies therefore try to purchase backup media and
create a schedule to automate of the backup process, thus not
requiring anyone to be physically present to change the tape media.
Incremental backups are much faster because they back up only
those files that have been changed since the last backup. The Archive
attribute switches on when a file is modified. An incremental backup backs up the file and then removes the attribute so that the file
will not be backed up again unless it is changed the next day. A
combination of incremental and full backups usually results in four
to six incremental tape sets and one full tape set each week. If the
drives fail, the administrator must restore the last full backup set, as
well as all the incremental backups performed since the drive failure.
This process obviously is considerably slower than a backup scheme
in which a full backup is performed every night.
Differential backups are similar to incremental backups except that
they do not reset the Archive attribute, which means that each backup during the week backs up all files changed since the last full
backup. A full backup once a week (generally Friday or Saturday)
and differentials every other day means that theoretically only two
tapes are needed in case of failure: the last full backup and the last
differential (see Figure 9.1).
Keeping a log of all backups is important. Most backup utilities can
generate a backup log. Microsoft recommends that you make two
copies of the backup log—store one with the backup tapes and keep
one at the computer site. Always test your backup system before you
trust it. Perform a sample backup, restore the data, and check the
data to be sure it is identical to the original.
DISASTER RECOV E RY
321
14.919-x CH09.i
322
8/28/98 1:38 PM
C h a p te r 9
Page 322
D I S A S TE R R E C OV ERY
FIGURE 9.1
An ideal backup scheme implements a schedule of different backup types.
Daily Full
Backup
M
T
W
T
F
M
T
W
T
F
M
T
W
T
F
Daily Incremental
Plus Weekly Full
Daily Differential
Plus Weekly Full
EXAM
TIP
Used to recover from a Thursday failure
Backup Utilities A number of
other vendors also offer backup
software—such as Arcadia’s
BackupExec or Cheyenne’s
ArcServe—that include additional
features, and in many cases, these
are a very wise investment. For the
test, though, remember that only
the Microsoft Backup utility will be
covered.
You can attach a tape drive directly to a single server, or you can
back up several servers across the network at once. Backups over the
network are convenient for the administrator, but they can produce
considerable network traffic. You can reduce the effects of this extra
traffic if you place the computer attached to the tape drive on an isolated network segment and connect it directly to secondary network
interface cards on each of the servers.
A final point has to do with how long you wish to keep your stored
data on the tape media. There is no correct time length. Some companies overwrite old tapes on a weekly basis, while others can keep
their tape backups indefinitely. The correct time length depends on
how important old data is to your firm. There is no simple answer to
this question.
Uninterruptible Power Supply
An Uninterruptible Power Supply (UPS) is a special battery (or
sometimes a generator) that supplies power to an electronic device in
14.919-x CH09.i
8/28/98 1:38 PM
Page 323
C hapter 9
DISASTER RECOV E RY
323
the event of a power failure. UPSs are commonly used with network
servers to prevent a disorderly shutdown by warning users to log
out. After a predetermined waiting period, the UPS software performs an orderly shutdown of the server. Many UPS units also regulate power distribution and serve as protection against power surges.
Remember that in most cases a UPS generally does not provide for
continued network functionality for longer than a few minutes. A
UPS is not intended to keep the server running through a long
power outage, but rather is designed to give the server time to do
what it needs to before shutting down. This can prevent the data
loss and system corruption that sometimes results from sudden shutdown. Some networks also have UPSs connected to their hubs and
routers as well, giving administrators remote access to the servers so
they can perform shutdown tasks in the event of a power outage.
When purchasing a UPS for a server, note that they come in many
varieties (see Figure 9.2). As noted earlier, the UPS is really just a
battery backup. Just like a car battery, the more powerful it is, the
more expensive it is. Prices run from the hundreds to many thousands of dollars. Before you buy, know how many servers you will be
running off the UPS and how much time they need to shut down
properly. One of the most popular UPS manufacturers is APC
(American Power Conversion), a company that offers a full line of
power supply and UPS products.
Power
Source
UPS
Server
Power
Source
Larger UPS
Server
Server
Server
FIGURE 9.2
A large UPS can service numerous components
at once.
14.919-x CH09.i
324
8/28/98 1:38 PM
C h a p te r 9
Page 324
D I S A S TE R R E C OV ERY
In summary, a UPS enables a server to shut down gracefully. This in
turn allows time for files to be saved, and corruption of data to be
kept to a minimum. Backups mainly provide a quick method for
system recovery. They require a long and tedious restoration process
that can cost your company dearly in lost revenue and productivity.
The following sections therefore examine some methods of minimizing—or even preventing—downtime in the event of a drive failure.
RECOVERING
FROM
SYSTEM FAILURE
Next to data security, keeping the network up and running properly
is the most crucial day-to-day task of an administrator. The loss of a
hard drive, even if not disastrous, can be a major inconvenience to
your network users and may cost your organization in lost time and
money. Procedures for lessening or preventing downtime from single
hardware failures should be implemented. Disk configurations that
enable this sort of protection are called fault-tolerant configurations.
It should be noted that fault-tolerant configurations are not designed
as a replacement for system tape backups.
NOTE
Implementing a Fault-Tolerant Design
Balancing Needs and Costs When
developing a fault-tolerance scheme,
remember that you must balance the
need for rapid recovery from a failure
against cost. The basic theory behind
fault-tolerant design is hardware
redundancy, which translates into
additional hardware expenses. Also,
remember that the greater the level of
redundancy, the greater the complexity
involved in the implementation.
Connecting network components into a fault-tolerant configuration
ensures that one hardware failure doesn’t halt the network. You can
achieve network fault tolerance by providing redundant data paths,
redundant hubs, and other such features. Generally, however, the
data on the server itself—its hard drives—is the most crucial.
Using RAID
A vital tool for protecting a network’s data is the use of a Redundant
Array of Inexpensive Disks (RAID). Using a RAID system enables
you to set up the best disk array design to protect your system. A
RAID system combines two or more disks to create a large virtual
disk structure that enables you to store redundant copies of the data.
In a disk array, the drives are coordinated into different levels of
RAID, to which the controller card distributes the data.
RAID uses a format of splitting data among drives at the bit, byte,
or block level. The term data striping refers to the capability of
14.919-x CH09.i
8/28/98 1:38 PM
Page 325
arranging data in different sequences across drives. Demonstration
of data stripping are shown in Figure 9.3. Microsoft calls this disk
stripping.
Your input in designing the most reliable drive setup for your network is an important responsibility. You must choose the best RAID
implementation level to meet your users’ requirements in data
integrity and cost. Seven levels of RAID are available on the market
today: 0, 1, 2, 3, 4, 5, 6, and 10. A higher number isn’t necessarily
indicative of a better choice, so you must select the best level for
WA R N I N G
C hapter 9
DISASTER RECOV E RY
Fault-Tolerance Does Not Replace
Backups! A fault-tolerant disk
scheme is used only to speed
recovery time from a hardware
fault. None of these RAID levels is
intended to be a replacement for
regular tape backups.
BYTE
1
2
3
4
5
6
7
8
BITS
1
2
7
DRIVE 1
8
DRIVE 4
3
4
5
DRIVE 2
6
CPU with
Disk Array
Controller
DRIVE 3
BASIC DATA STRIPING
BYTE
1
2
3
4
5
6
1 2 3 4
7
8
7 8 1 2
DRIVE 1
DRIVE 4
3 4 5 6
5 6 7 8
DRIVE 2
DRIVE 3
REDUNDANT DATA STRIPING
CPU with
Disk Array
Controller
FIGURE 9.3
Data striping arranges data in different
sequences across drives.
325
14.919-x CH09.i
326
8/28/98 1:38 PM
C h a p te r 9
Page 326
D I S A S TE R R E C OV ERY
your environment. The following paragraphs present a brief discussion of some of these available levels, notably RAID 0, 1, and 5,
which Windows NT Server supports. Windows NT Workstation
supports only RAID 0, and Windows 95 is not able to use any
RAID levels at all.
RAID 0
RAID 0 uses data striping and block interleaving, a process that
involves distributing the data block by block across the disk array in
the same location across each disk. Data can be read or written to
these same sectors from either disk, thus improving performance.
RAID 0 requires at least two disks, and the striped partitions must
be of the same size. Note that redundancy of data is not provided in
RAID 0, which means that the failure of any single drive in the array
can bring down the entire system and result in the loss of all data
contained in the array. RAID 0 is supported in Windows NT Server
and Windows NT Workstation, but not in Windows 95. In short,
RAID 0 does not provide any fault tolerance, just faster disk drive
performance.
RAID 1
In RAID 1, drives are paired or mirrored: Each byte of information
is written to two identical drives. Disk mirroring is defined as two
hard drives—one primary, one secondary—that use the same disk
channel (controller cards and cable), as shown in Figure 9.4. Disk
mirroring is most commonly configured by using disk drives contained in the server.
Disk mirroring is called disk duplexing when a separate drive controller is added for each drive. Duplexing, which is covered later in
this chapter, is a form of mirroring that enables you to configure a
more robust hardware environment.
Mirroring does not provide a performance benefit such as RAID 0
provides. You can use mirroring, however, to create two copies of the
server’s data and operating system, which enables either disk to boot
and run the server. If one drive in the pair fails, for instance, the
other drive can continue to operate. Disk mirroring can be expensive, though, because it requires 2GB of disk space for every 1GB
you want to mirror. You also must make sure that your power source
has enough wattage to handle the additional devices. Mirroring
requires two drives, and the mirrored partitions must be of the same
14.919-x CH09.i
8/28/98 1:38 PM
Page 327
C hapter 9
DISASTER RECOV E RY
327
FIGURE 9.4
In disk mirroring, two hard drives use the same
disk channel.
DRIVE 0
(DATA IS WRITTEN AND
READ HERE FIRST)
DISK
CONTROLLER
(CHANNEL 0)
DRIVE 1
(DATA IS WRITTEN AND
READ HERE SECOND)
size. Windows NT Server supports mirroring, but Windows NT
Workstation and Windows 95 do not.
Remember that mirroring is done for fault-tolerant, not performance reasons. With this said, it should be noted that a Windows
NT machine running a mirror set runs at about normal speed. It
may exhibit a degradation if only one controller card is shared by
the two hard drives. The controller must make each write twice,
once for each drive. On the other hand, a mirrored hard drive set
can produce marginal performance gains reading from the set
because either drive can satisfy the read. For the best of both worlds,
though, consider RAID 5.
R A I D L E V E L 5 M O S T POP ULA R S CHE ME
RAID Levels Supported in Windows NTRAID 2, 3, 4, and 5 are all
versions of striping that incorporate similar fault-tolerant designs.
Microsoft chose to support only RAID 5 striping in Windows NT
Server. As the numbering scheme would imply, this is the newest
revision of the four and is the most popular fault-tolerance scheme
in use today. Level 5 requires less disk space than mirroring and
has performance gains over other striping methods. As with mirroring, RAID level 5 is not available in Windows NT Workstation or
Windows 95.
14.919-x CH09.i
328
8/28/98 1:38 PM
C h a p te r 9
Page 328
D I S A S TE R R E C OV ERY
RAID 5
RAID 5 uses striping with parity information written across multiple
drives to enable fault tolerance with a minimum of wasted disk
space. This level also offers the advantage of enabling relatively efficient performance on writes to the drives, as well as excellent read
performance.
Striping with parity is based on the principle that all data is written
to the hard drive in binary code (ones and zeros). RAID 5 requires
at least three drives because this version writes data across two of
them and then creates the parity block on the third. This writing of
data and the parity bit is spanned across all drives being used. If the
first byte is 00111000 and the second is 10101001, then the system
computes the third by adding the digits together using this system:
1+1=0, 0+0=0, 0+1=1, 1+0=1
The sum of 00111000 and 10101001 is 10010001, which is written
to the third disk. This process would continues as the next parity bit
is written to the first drive, and the data to the second and third. On
the third round, the parity bit is written to the second drive and the
data to the first and third drive. Then this cycle repeats itself.
If any of the disks fail, the process can be reversed and any disk can
be reconstructed from the data and parity bits on the other two. See
Figure 9.5 for an illustration of the process. Recovery includes
replacing the bad disk and then regenerating its data through the
Disk Administrator. A maximum of 32 disks can be connected in a
RAID 5 array under Windows NT.
Choosing a RAID Level
When implementing a disk scheme, you have some options to consider. First, you must decide whether you are interested in performance gains (RAID 0), often used by read-only databases loaded
from a CD-ROM, or data redundancy (RAID 1 or 5), often
required by systems that need real-time access to continually changing data, such as a scheduling system. Mirroring (RAID 1), for
instance, enables the fastest recovery but results in a 50% loss of disk
space. Likewise, striping with parity (RAID 5) is more economical
but requires at least three physical disks and therefore provides more
points of potential hardware failure. RAID 0 makes sense when the
data change little or not at all and are available from CD-ROM or
14.919-x CH09.i
8/28/98 1:38 PM
Page 329
C hapter 9
DISASTER RECOV E RY
329
FIGURE 9.5
Disk 3
Disk 2
Disk 1
+
=
1 0 1 1 0 0 1 0
0 0 1 1 1 0 0 1
1 0 0 0 1 0 1 1
Data written
to disk 1
Data written
to disk 2
Parity Data
If disk 2 fails, the system is able to reconstruct the
information on it by using the parity data…
Disk 3
Disk 1
+
1 0 1 1 0 0 1 0
=
1 0 0 0 1 0 1 1
0 0 1 1 1 0 0 1
other types of backup storage. In Windows NT, all the RAID levels
are supported as software implementations of RAID, but you can
implement hardware versions as well.
Most network administrators prefer the RAID 5 solution, at least on
larger servers with multiple drive bays. Because this level is a hybrid
of striping and mirroring, it enables greater speed and more redundancy. Mirroring, however, offers the advantage of working well
with non-SCSI hardware, because some older machines accommodate two IDE drives only, and is common as a fault-tolerant option
on smaller, non-dedicated servers. Striping without parity should be
reserved for workstations and servers on which speed considerations
are paramount and possible downtime is an acceptable risk. See
Figure 9.6 for a graphical comparison.
Disk Duplexing
In the event of disk channel failure (by a controller card or cable),
access to all data on the channel stops and a message appears on the
file server console screen (if your users don’t let you know about it
first). Even though drives can be mirrored, all disk activity on the
mirrored pair ceases if the mirrored drives are connected to the same
disk controller.
In this example, if Disk 2 fails, the system can
reconstruct the information on it using the parity data.
14.919-x CH09.i
330
8/28/98 1:39 PM
C h a p te r 9
Page 330
D I S A S TE R R E C OV ERY
FIGURE 9.6
RAID 0 - Disk Striping
Different RAID levels offer their own unique
capabilities.
+
=
Requires at least two disks
Configured for performance gain, NOT FAULT TOLERANT
RAID 1 - Disk Mirroring
=
Fault Tolerant
Wastes 50% of disk space
Can slow down the system on extensive writes.
RAID 5 - Disk Striping with Parity
+
=
Fault Tolerant
More efficient in disk usage than mirroring
Performance aided by striping, slowed by writing parity
End result is moderate write performance, fast reads
Disk duplexing performs the function of simultaneously writing data
to disks located on different channels. As Figure 9.7 illustrates, each
hard disk in a duplexed pair connects to a separate hard disk controller. This figure shows a configuration in which the drives are
housed in separate disk subsystems. Each subsystem also has a separate power supply. Disk duplexing offers a more reliable setup than is
possible with mirroring because a failure of one disk drive’s power
supply doesn’t disable the server. Instead, the server continues to
work with the system that remains under power.
Working on the same channel is analogous to going to a baseball
game when only one gate into the stadium is open. You can enter
or exit through only one gate (channel) at the stadium (file server),
and the crowd (data) can get backed up on both sides. If more
than one gate (another channel) is open, though, the crowd (data)
doesn’t become backed up on both sides of the fence (file server or
workstation). This is why disk duplexing, which uses a separate
14.919-x CH09.i
8/28/98 1:39 PM
Page 331
C hapter 9
DISASTER RECOV E RY
331
FIGURE 9.7
Disk duplexing simultaneously writes data to
two disks located on different controller cards.
DISK
CONTROLLER 0
(CHANNEL 0)
DISK
CONTROLLER 1
(CHANNEL 1)
DRIVE 1
DRIVE 2
DRIVE 0
DRIVE 0
DISK SUBSYSTEM 1
DISK SUBSYSTEM 0
adapter card for each disk, has faster reads and writes than disk mirroring.
Duplexing protects information at the hardware level with duplicate
channels (controller cards and cables) and duplicate hard drives.
Mirroring uses one controller card and two hard drives. The point
of failure for this setup is primarily the controller card or the cable
connecting the drives to the controller card. Disk duplexing uses
two controller cards and a minimum of one drive per controller
card. The point of failure is reduced with duplicate hardware.
T H I R D - PA R T Y OP TI ONS
A number of different vendors also offer RAID protection at the
hardware level on their server products. This protection is independent of the operating system, so if you really feel that RAID 5 on
your Windows 95 workstation is a necessity, these software vendors might have a solution for you.
14.919-x CH09.i
332
8/28/98 1:39 PM
C h a p te r 9
Page 332
D I S A S TE R R E C OV ERY
The previous sections examined a number of different disk configurations. Exercise 9.1 shows how you implement these RAID levels
and other disk configuration options in Windows NT. The primary
program for managing disk storage resources is the Disk Administrator, a tool that is generally usable only by members of the Administrators or Server Operators groups.
OTHER FAULT-TOLERANCE
MECHANISMS
Two other forms of fault tolerance exist on the market today. One of
these is known as server mirroring while the other is a hardware solution known as a super server.
Server mirroring refers to having one server completely mirrored in
all forms to another server. This means that if Server A goes down
for any reason whatsoever, such as a failed hard drive, failed network
card, or even a blown motherboard, the mirrored Server B takes over
the duties of Server A. This type of fault tolerance is offered by
Microsoft in their Microsoft Cluster Server product.
A second option on fault tolerance is a super server. A super server is
a hardware solution offered by several different hardware manufacturers. The idea behind a super server is that almost any piece of
equipment can be changed on the super server without shutting
down the server. This can mean that the super server can have hot
swappable components such as hard drives, CPUs, and even RAM.
14.919-x CH09.i
8/28/98 1:39 PM
Page 333
C hapter 9
DISASTER RECOV E RY
C A S E S T U DY : I M P L E M E N T I N G F A U LT T O L E R A N C E
C O R P O R AT E N E T W O R K
FOR A
ESSENCE OF THE CASE
A N A LY S I S
The essence of the case is as follows:
This analysis is presented in two parts: Backup
solutions and fault-tolerance options.
• The company works on a small LAN with
one server.
• A backup strategy is needed.
• Restore time should preferably not go
over two hours.
• The insurance company wishes to be
made aware of the costs and benefits of
different fault-tolerance options.
SCENARIO
You are meeting with the branch manager of an
insurance firm. This is a large insurance firm that
employs fifty insurance brokers. They have just
changed over from their old paper methods of filing insurance claims and contracts to using computers on a LAN.
The LAN is composed of 30 workstations and
one file server. All data is stored on the file server which has a 3GB drive. Nothing of importance
is stored on the workstations. Downtime should
be kept to a 2 hour maximum, because brokers
can still sell insurance using the software on
their PCs, but if the server is down, they cannot
retrieve any client information.
This insurance company is now very interested in
designing a backup strategy and possibly
installing some form of fault tolerance for the
branch. Your job is to provide a backup solution
and give a general cost analysis of different
fault-tolerance methods.
333
Backup Solutions
This company has a server with a very small hard
drive of only three gigabytes in size. By today’s
standard this is not very large. Because a single
backup tape can accommodate this size of a
hard drive, a simple solution would be to do a full
backup. No one needs to be around to switch
tapes, and the backup process should not take
more than a couple of hours. Also, restoring can
be done much more quickly from a full backup
than from a differential or incremental backup.
One could also make the argument that a differential backup should be done each night, with a
full backup being done every week. This is often
done when companies have such a large amount
of data being backed up that one tape is not sufficient to do a full backup or when there are not
enough hours during the night to accomplish a
full backup. Because this company’s data storage is not that large, and there is easily enough
time in the evening to do a full backup, there is
no reason why a full backup should not be done
each night.
Fault Tolerance
This insurance company has several options for
fault tolerance. These range from disk mirroring
to clustered super servers. Almost all the faulttolerance options can be done together.
continues
14.919-x CH09.i
334
8/28/98 1:39 PM
C h a p te r 9
Page 334
D I S A S TE R R E C OV ERY
C A S E S T U DY : I M P L E M E N T I N G F A U LT T O L E R A N C E
C O R P O R AT E N E T W O R K
FOR A
continued
UPS
A UPS should be purchased. The cost could be
only a few hundred dollars, but the benefit of the
UPS in preventing data corruption due to a power
failure would more than offset this cost.
Disk Mirroring/Disk Duplexing
This option can be done. It would mean the purchase of an additional 3GB drive. If disk duplexing was done, it would also mean the purchase
of a disk controller. If either of the drives failed
(or controller cards if you were doing duplexing),
the other drive could take over. This would prevent downtime for the company. This option
would probably cost under $1000.
Disk Striping with Parity
This option is feasible. It would mean purchasing
at least two more drives. If any of the drives
failed, the data could be regenerated from the
striped parity bit. This option would also cost
probably under $1000. One thing to be aware of
is that because the costs of hard drives are relatively low nowadays, disk striping with parity, in
the case of the 3GB of storage, is really not
much more expensive than disk mirroring, but
would also provide an extra 3GB of disk storage
space (original 3GB + 2 more 3GB drives equals
9GB, with 1/3 of the drive space used for the
parity bit).
Server Clustering
Server clustering is also an option. This option
would cost the equivalent of the price of all the
hardware and software of the original server, plus
the price of the clustering software. The benefit
is that if any component fails in either server, the
other clustered server can take over the functions of the downed server.
Super Server
This option would be the most expensive. It
would cost well into the tens of thousands of dollars. The benefit is that if any component fails, a
backup component can take over. In addition,
faulty components can be replaced without downing the server.
Based on the fact that the company can afford
several hours of downtime, the options for the
super server and clustering would more than likely be voted down, because they exceed the
requirements of the insurance company. The
options of disk mirroring and disk
striping/duplexing with parity are both contenders, and probably should be recommended
based upon whichever option is cheapest.
14.919-x CH09.i
8/28/98 1:39 PM
Page 335
C hapter 9
DISASTER RECOV E RY
CHAPTER SUMMARY
This chapter examined a number of options open to an administrator looking to provide data security and hardware redundancy for
the network. Through the use of a regular backup plan, the installation of a UPS, and the implementation of a fault-tolerant disk
scheme, you can help to ensure that your network will run as efficiently and safely as possible. Remember that there is no particular
formula to use here; rather, you should follow a process of weighing
costs against benefits. In the end, you want to provide the highest
degree of safety for your critical data that you can achieve given your
budget.
KEY TERMS
• Full backup
• Incremental backup
• Differential backup
• Daily copy
• Copy
• Uninterruptible Power Supply
(UPS)
• RAID
• Data striping
• Disk striping
• Block interleaving
• Disk mirroring
• Disk duplexing
• Disk striping with parity
• Server mirroring
• Super server
335
14.919-x CH09.i
336
8/28/98 1:39 PM
C h a p te r 9
Page 336
D I S A S TE R R E C OV ERY
WA R N I N G
A P P LY Y O U R L E A R N I N G
Making Changes in Disk
Administrator Remember that
changes made to your disk configuration can have a serious effect on
the system. Do not make any
changes in Disk Administrator
unless you have carefully planned
them previously!
Exercises
9.1
Exploring Windows NT’s Disk
Administrator
Objective: Explore the options available through Disk
Administrator, such as establishing and breaking mirrored drives and creating or regenerating stripe sets
with parity.
Estimated time: 10 minutes
To complete exercise 9.1, log on to a Windows NT 4.0
server or workstation with an account that has administrative authority. The server or workstation used can be
a production machine—no changes will actually be
made to the computer’s configuration during this exercise if the steps in this exercise are followed.
1. Click Start, Programs, Administrative Tools.
Then choose Disk Administrator. If this is the
first time the application is run, or if disks have
been added to the system, you will be asked for
permission to write a signature block to the disk.
If this message appears, click on Yes.
2. Observe the Disk Administrator window and
maximize it if it is not already in this state. The
configuration of the disk or disks on your
machine is displayed.
3. Click one of the partitions on your screen. A dark
black line appears around the partition, indicating that the partition is selected. Right-click on
the partition and observe the available menu
choices in the context-sensitive menu. Note that
you can format the partition, delete the partition,
change its logical drive letter, or examine its properties. If the disk is removable, the Eject option is
also available.
4. Click Partition in the Menu bar and examine the
choices. Most of the choices are unavailable, but
they include Create Volume Set and Create Stripe
Set. You also can change your active partition in
this Menu bar.
5. Click Fault-tolerance on the Menu bar (Windows
NT Server only) and observe that this menu
enables you to establish and break mirrored drives, as well as to create or regenerate stripe sets
with parity.
6. Feel free to explore further, and when you are finished examining the menus and options, close
out of the Disk Administrator by clicking
Partition, Exit. If you are asked to commit or save
your changes, click Cancel.
Review Questions
1. What method of fault tolerance uses two controller cards on two separate hard drives?
2. What are the three backup methods that are not
Microsoft NT NTBACKUP-specific?
14.919-x CH09.i
8/28/98 1:39 PM
Page 337
C hapter 9
DISASTER RECOV E RY
337
A P P LY Y O U R L E A R N I N G
Exam Questions
1. What is an incremental backup?
A. Backs up parts of the specified file that have
changed since the last backup
B. Backs up and marks only those files that have
changed since they were last backed up
C. Backs up the files that have changed since
they were last backed up but doesn’t mark
them as being backed up
D. Backs up the files that have changed over the
course of a specified time period
2. What is a differential backup?
A. Backs up files that have changed since the last
backup and doesn’t mark the files as having
been backed up
B. Backs up files that have changed since the last
backup and marks the files as having been
backed up
C. Copies all files that have been modified within a specific time period and marks them as
having been backed up
D. Copies all files that have been modified within a specified time period and doesn’t mark
them as having been backed up
3. What is the best way to reduce the effects of extra
traffic caused by a network backup?
C. Place the computer attached to the tape drive
on an isolated network segment
D. Back up the servers in ascending order of the
size of the backup
4. What does UPS stands for?
A. Unintentional Packet Switch
B. Unfamiliar Password Sequence
C. Unknown Polling Sequence
D. Uninterruptible Power Supply
5. How does RAID Level 5 operate?
A. Uses bit interleave data striping
B. Uses block interleave data striping
C. Doesn’t use data striping
D. Provides parity-checking capabilities
6. How does RAID Level 1 operate?
A. Uses bit interleave data striping
B. Uses block interleave data striping
C. Doesn’t use data striping
D. Provides parity-checking capabilities
7. What is the difference between disk mirroring
and disk duplexing?
A. Disk mirroring is more reliable.
B. Mirrored disks share the same disk channels.
A. Attach the tape drive directly to one of the
servers
C. Duplexed disks share the same disk channels.
B. Back up each server to a nearby server
D. There is no difference.
14.919-x CH09.i
338
8/28/98 1:39 PM
C h a p te r 9
Page 338
D I S A S TE R R E C OV ERY
A P P LY Y O U R L E A R N I N G
8. True or False: Implementing a RAID system
eliminates the need for tape backup.
C. 16
D. None of the above
A. True
B. False
9. What is the minimum number of disks needed to
configure a stripe set with parity on Windows
NT Server?
A. Two
B. Three
C. Four
D. Seven
10. RAID 5 is a term that describes which of the following?
A. A weekday backup strategy for enterprise networks
B. A fault-tolerant disk configuration
13. A corporate network running Windows NT
would like to increase the fault tolerance on its
systems.
Primary objective: The system cannot have a
drive failure cause the system to become inaccessible.
Secondary objective: The server on the network
needs to be able to shut down properly in the
event of a power failure.
Secondary objective: Data access speed is not critical.
Suggested Solution: You install three hard drives,
set up disk striping, and add a UPS to the server.
A. This solution meets the primary objective and
both secondary objectives.
C. An NDIS-compatible SCSI controller
B. This solution meets the primary objective and
one secondary objective.
D. Data backup through directory replication
C. This solution meets the primary objective.
11. What is the maximum number of disks in a
stripe set for NT?
A. 2
B. 16
C. 32
D. Limited only by hardware
12. What is the maximum number of drives supported in a mirror set?
A. 2
B. 4
D. This solution does not meet the primary
objective.
14. A corporate network running Windows NT
would like to increase the fault tolerance on its
systems.
Primary objective: The system cannot have a
drive failure cause the system to become inaccessible.
Secondary objective: The server on the network
needs to be able to shut down properly in the
event of a power failure.
14.919-x CH09.i
8/28/98 1:39 PM
Page 339
C hapter 9
DISASTER RECOV E RY
339
A P P LY Y O U R L E A R N I N G
Secondary objective: Data access speed is not critical.
Suggested Solution: You install three hard drives
and set up Disk Striping with Parity.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objectives.
C. This solution meets the primary objective.
D. This solution does not satisfy the primary
objective.
15. A corporate network running Windows NT
would like to increase the fault tolerance on its
systems.
Primary objective: The system cannot have a
drive failure cause the system to become inaccessible.
Secondary objective: The server on the network
needs to be able to shut down properly in the
event of a power failure.
Secondary objective: Data access speed is not critical.
Suggested Solution: You install two hard drives,
set up disk duplexing, and add a UPS to the
server.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objectives.
C. This solution meets the primary objective.
D. This solution does not satisfy the primary
objective.
Answers to Review Questions
1. Disk duplexing uses two controller cards, one for
each disk. Disk duplexing is the same as disk mirroring, but disk mirroring uses only one controller card. See the section titled “Disk
Duplexing.”
2. The three methods of backup are full, incremental, and differential backups. Full back up backs
up all data. An incremental backup backs up only
the data that has changed since the last incremental backup. Incremental backups remove the
archive bit on a file. The third option is differential backup. A differential backup does not
remove the archive bit on a file, and backs up all
data since the last full backup. See the section
titled “Backup.”
Answers to Exam Questions
1. B. A is incorrect because the entire file is always
backed up. C is incorrect because this is a differential backup. D is incorrect because files are not
backed up over “the course of a specified time.”
See the section titled “Backup.”
2. A. B is incorrect because the files are not marked
in a differential backup. C is incorrect because no
files are copied, and nothing is marked as “backed
up.” D is incorrect because files are not backed
up due to a time interval. See the section titled
“Backup.”
3. C. A is incorrect because a tape drive attached to
the computer to back up that computer is not a
network backup. B is incorrect because it does
not reduce network traffic. D is incorrect because
backing up according to server size does not
14.919-x CH09.i
340
8/28/98 1:39 PM
C h a p te r 9
Page 340
D I S A S TE R R E C OV ERY
A P P LY Y O U R L E A R N I N G
reduce the network traffic. See the section titled
“Backup.”
11. C. 32 is the maximum disks handled in a stripe
set for NT. See the section titled “RAID 5.”
4. D. All answers but D are made up. See the section titled “Uninterruptible Power Supply.”
12. A. Mirror sets contain only 2 disks. See the section titled “Disk Duplexing.”
5. D. All answers but D do not apply to RAID 5.
See the section titled “RAID5.”
13. D. Disk Striping does not provide any fault tolerance. It yields faster reads when accessing the
data. Adding a UPS does meet both secondary
objectives. The primary objective is not met by
implementing Disk Striping (RAID 0). Disk
striping with parity and disk duplexing would
meet the required solution. See the sections titled
“Implementing a Fault-Tolerant Design” and
“Uninterruptible Power Supply.”
6. C. A and B have nothing to do with RAID 1. D
is a function of RAID 5. See the section titled
“RAID 1.”
7. B. Disk duplexing is disk mirroring, but with
each hard drive on a separate controller card. See
the section titled “Disk Duplexing.”
8. B. RAID systems never replace the need for a
backup. RAID systems provide a short term solution for drive failures. See the section titled
“Implementing a Fault-Tolerant Design.”
9. B. Stripe sets with parity need a minimum of
three disks. Disk mirroring requires only two
disks. See the section titled “RAID 5.”
10. B. RAID 5 is a fault-tolerant “Redundant Array
of Inexpensive Disks.” See the section titled
“RAID 5.”
14. B. Disk striping with parity meets the primary
objective and second secondary objective. Because
the proposed solution contains nothing about
power outages, the first secondary objective is not
met. See the sections titled “Implementing a
Fault-Tolerant Design” and “Uninterruptible
Power Supply.”
15. A. All objectives are met with this solution. See
the sections titled “Implementing a FaultTolerant Design” and “Uninterruptible Power
Supply.”
Suggested Readings and Resources
1. Siyan, Karanjit S. Windows NT Server 4
Professional Reference. New Riders, 1996.
3. Casad, Joe. MCSE Training Guide: Windows
NT Server 4. New Riders, 1997.
2. Heywood, Drew. Inside Windows NT Server 4.
New Riders, 1997.
4. Sirockman, Jason. MCSE Training Guide:
Windows NT Server 4 Enterprise. New Riders,
1997.
15.919-x Part3
8/28/98 1:46 PM
Page 341
P A R T
IMPLEMENTATION
10 Managing and Securing a Microsoft Network
11 Monitoring the Network
III
15.919-x Part3
8/28/98 1:46 PM
Page 342
16.919-x CH10.i
8/28/98 1:47 PM
Page 343
OBJECTIVES
Chapter 10 targets the following objectives in the
“Implementation” and the “Standards and
Terminology” sections of the Networking Essentials
exam:
Choose an administrative plan to meet specified
needs, including performance management,
account management, and security
. Network security can vary depending on the network operating system being used. Because of the
existence of different security models, it is important to understand the different administrative
models that exist. One model may be ideal for one
situation, but impractical for another. This chapter
analyzes these issues and explains the various
administrative models in order to address the issues
of performance, account management, and security.
To master this exam topic, pay particular attention
to the differences between Workgroup and Domain
administrative models. This objective is addressed
throughout the entire chapter.
Compare user-level security with access permission assigned to a shared directory on a server
. This exam objective is designed to encourage you
to know the different permissions available to
assign to users or groups on a shared directory.
C H A P T E R
10
Managing and
Securing a Microsoft
Network
16.919-x CH10.i
8/28/98 1:47 PM
Page 344
OUTLINE
Resource Sharing Basics
346
Resources
346
Sharing
347
Users
347
Groups
347
Security
348
General Network Administrative
Models
Implementing Security on Windows NT
364
Creating and Assigning Permissions
to a Shared Folder in Windows NT
365
Assigning File-Level Permissions on
an NTFS Partition
366
Printer Sharing with Windows NT
367
348
Implementing Security on Windows 95
368
Workgroup Model
Windows 95
Windows NT
349
350
350
Share-Level Security on Windows 95
368
User-Level Security on Windows 95
369
Bindery-Based Model
352
Printer Sharing with Windows 95
371
Domain Model
354
Directory Services Model
356
Additional Administrative Tasks
372
Auditing
372
Handling Data Encryption
372
358
Handling Virus Protection
373
User Accounts
358
Securing Equipment
373
Groups
Global
Local
360
360
360
Permissions
363
Rights
363
Managing User Accounts and Groups
Using Windows NT
Chapter Summary
376
16.919-x CH10.i
8/28/98 1:47 PM
Page 345
S T U DY S T R AT E G I E S
. This chapter addresses two exam topics. In
order to study for these topics, pay particular
attention to the differences in what Windows 95
and Windows NT computers offer in terms of
security. Be aware that Windows NT security is
much more complex than that offered by
Windows 95.
. Both Windows 95 and Windows NT offer share
level security. Be aware of the differences
between share level security on Windows NT
and Windows 95, as well as the difference
between file level and share level security that
is offered by Windows NT.
16.919-x CH10.i
346
8/28/98 1:47 PM
C h a p te r 1 0
Page 346
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
INTRODUCTION
EXAM
TIP
In the preceding chapters, the process of establishing a physical connection between the machines on your network and installing the
drivers and services necessary to enable network communication was
examined. With these initial considerations out of the way, the next
step is to begin organizing and controlling the manner and scope of
network usage. This chapter deals with the process of implementing
resource sharing, with the main focus being the administration of a
Microsoft network.
Resource Security in Windows NT
and Windows 95 Pay particular
attention to how Windows NT and
Windows 95 perform resource
security. Be prepared to explain the
differences between these two
major models.
The process of implementing resource sharing will be presented in
the following order: First, a general overview of some key resource
terms is presented. From this perspective, several different administrative models will be presented and contrasted. You should focus on
the administrative models supported by Windows NT and Windows
95. After you have this background, file security and then print
security will be analyzed from both Windows NT and Windows 95
perspectives. The final area of discussion will focus on some additional administrative tasks that should be performed on a network
RESOURCE SHARING BASICS
Microsoft uses very specific terms to describe elements of its networking structure, and as such, a good understanding of these terms
is essential. The five most basic terms that you must understand are
resources, sharing, users, groups, and security.
Resources
The first concept to be discussed is a resource. A resource is essentially any component that you would like to use on the network. This
could be as simple as a file on another machine, to a printer located
at the end of the hall, to even a certain task available by a specific
program. The two key resources detailed in this chapter are data files
and printers, but in theory, a resource can be any information or
device relating to the network. Without networking, a resource can
be accessed only by physically sitting at the machine on which the
Page 347
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
resource is installed. This would mean that you could only access a
local file or a local printer. The creation of a networking structure
grants you the capability to use a server computer to share resources
with others at remote client machines.
What Is a Server? Remember that
the term “server” does not refer
exclusively to Windows NT Server. The
term is actually a generic reference to
any computer that provides resources
or services to other machines on the
network. As such, Windows 95,
Windows NT Workstation, and
Windows for Workgroups all are capable of performing the basic functions
of a server in certain instances, such
as file and printer sharing.
The Network Browser The browser,
or “Network Neighborhood” as it is
also called in Windows 95 and
Windows NT, is a program that allows
a user to see resources on the network. These resources are seen if
they can advertise their services
using Microsoft’s browser service. The
ability to advertise shared directories
and printers is installed automatically
in Windows NT computers that are
networked, but only on Windows 95
computers that have the “File and
Print” service installed.
Sharing
This brings us to the second important concept: sharing. Only by
specifying that you want to grant others access to a resource—be it a
directory, a CD-ROM drive, or a printer—do you make the resource
available for use from remote computers and devices. A shared
resource is simply a resource whose owner has leveraged networking
to make it available for use by others. Some resources are not available until an administrator actually manually shares out the resource.
Some examples of these resources are files and printers. Other
resources are automatically shared out when installed. An example of
this is the ability to see a computer on the network when browsing
the network.
Users
A user is anyone who requests network resources. In most cases, you
assign a unique username and password to every individual on your
network. Users can be created on a number of operating systems,
including Windows NT, NetWare, and UNIX. Users cannot be created on Windows 95 or Windows for Workgroups because neither
of these operating systems have the capability of establishing a user
database. Both Windows 95 and Windows for Workgroups do
enable the creation of individualized profiles, but as you will see later
in the chapter, they must rely on another machine’s database to provide true user authentication, such as an Windows NT domain controller.
Groups
Groups are administrative units that are comprised of one or more
users with similar needs for network resources. Often users are
placed into groups, and resource access is managed on a group basis,
347
NOTE
8/28/98 1:47 PM
NOTE
16.919-x CH10.i
16.919-x CH10.i
348
8/28/98 1:47 PM
C h a p te r 1 0
Page 348
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
as opposed to an individual user basis. It is much easier to manage
five groups than five hundred users. Two types of groups exist on
Windows NT—local and global. These groups are key to efficient
security in the Microsoft model.
Security
The issue of security is one of the main focuses of this chapter.
Security is the process of giving “Rights” or “Permissions” to groups
or users, such that they can access resources on the network.
Different Network operating systems use different terms to describe
these types of security issues. Windows NT makes a distinction
between “Rights” and “Permissions.” The details of these differences
between these “Rights” and “Permissions” will be addressed in
greater detail later on in this chapter.
GENERAL NETWORK ADMINISTRATIVE
MODELS
Choose an administrative plan to meet specified needs, including
performance management, account management, and security.
There are numerous networking administrative models that perform
security to choose from. Most network operating systems follow or
include only one model, yet others allow you to pick from several
models. This section will discuss four commonly used network security models that are commonly used in networking today. These are:
á Workgroups
á Bindery-based
á Domains
á Directory services
These four classifications are not etched in stone, and are by no
means the only four options in existence. They do, however, represent the majority of the network security models found on the market today. When comparing these models, pay particular attention
to the sections on workgroups and domains, as these are the two
current models used by Microsoft networks.
8/28/98 1:47 PM
Page 349
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
Workgroup Model
One common security model used on small networks is the workgroup model. This administrative model is built into network operating systems such as Windows 95, Windows for Workgroups, and
Windows NT.
In a workgroup model, there is no centralized database or server that
stores user account information (see Figure 10.1). This type of security model is found on a peer-to-peer type network. In a workgroup
model, there are one or more machines that have a resource to share.
Assume this resource is a directory containing some files. In order to
allow other computers to access these files, the computers containing
these files in the directory must have a service running that allows
them advertise this sharing of resources.
NOTE
16.919-x CH10.i
349
The Apple Version of Workgroups
Apple computers also follow a model
similar to a workgroup model, but the
terminology used by Apple computers
is a “zone” as opposed to a “workgroup.” Multiple network segments
can be joined to form a single zone,
and a single segment can have multiple zones.
A workgroup is just a name associated with a group of computers.
Any computer, when installed, can be part of any workgroup they
wish. If none exists, you can install a computer to be part of a new
workgroup. The name of a workgroup is simply for organizational
purposes, such that when one uses a network browser, computers
that are part of the same workgroup will be clustered together. Often
workgroup names will be descriptive, such as ENGINEERING or
ACCOUNTING.
All machines belong to the workgroup: Sales
What is the password?
PC2
PC1
PC3
PC3 accessing a resource on PC1
PC4
FIGURE 10.1
A workgroup does not rely on a centralized user
account database.
16.919-x CH10.i
350
8/28/98 1:47 PM
C h a p te r 1 0
Page 350
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
There are two variations found within the Workgroup model. One
is a Windows 95 and Windows for Workgroups variation, while the
other is a Windows NT variation.
Windows 95
In order for users on other computers to access files or printers on
other computers, their computers must have a redirector (see
Chapter 1) installed that will allow them to connect to the advertising service. In Windows 95, the redirector is called “Client for
Microsoft Networks,” and the service that allows shared files to be
accessed over the network is called “File and printer sharing for
Microsoft networks.” Exercise 10.5 will give you hands-on practice
in installing and using these services.
When a resource is shared out on the network, this provides the
capability to allow users access to the resource, but there is no capability to give this access to the resource on a user-by-user, or groupby-group basis. This is because when sharing out a resource, you can
only specify a password to protect the resource. When anyone tries
to connect to this shared resource, they will be prompted for a password. If they type in the correct password, when prompted, they
will get access to the resource. If they do not know the password,
they will be denied access. This is very similar to “Ali Babba and the
40 thieves.” In order to get into the secret cave, Ali Babba needed to
know the password, which was “Open Sesame.” The workgroup
works on the same principle. Anyone who does know the password
will be allowed access; anyone who doesn’t will be denied.
This security model works well in small networks. As a network
grows, the use of passwords on every shared resource becomes cumbersome. There is also no method of controlling anyone from telling
others the password to your shared resources.
Windows NT
Windows NT Workstation and Server both have the ability to be
installed within a workgroup. This option is selected during the
installation of the software, but can, in most cases, also be done after
the computer is installed.
The Windows NT workgroup model works in a fashion similar to
that used by Windows 95, yet there is a major exception. Each
16.919-x CH10.i
8/28/98 1:47 PM
Page 351
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
Windows NT computer does contain a local database of user
accounts. In order to access a local Windows NT computer, you
would need to log on to the computer using a name and password
found in the local user account database. The contents of this local
user account database are not used with any other computers.
WH E N W I N D OW S 9 5 C A N OFFER US E R-LE V EL
S E C U R I TY
Windows 95 does have the ability to reference accounts on a
Windows NT server (in a workgroup or a domain) or Novell server
offering bindery services. This is known as implementing user level
security in Windows 95.
The ability for Windows 95 to provide this service does rely on the
presence of either a Windows NT domain controller, Windows NT
computer as part of a workgroup, or a Novell server with bindery
services being present. The user, when logging on to the Windows
95 computer, will have their logon credentials passed onto the
respective server for authentication.
In conclusion, when mixing Windows 95 and Windows NT (or Novell)
servers, it is possible for Windows 95 to offer user level security
and for Windows NT computers, as part of a workgroup, to share
their account databases.
Windows NT in a workgroup model has the capability to reference
users on a user-by-user basis when assigning security to shared
resources. The users a Windows NT computer can reference are only
the ones found within its own user account database. To have a security model, you would need to create all of the users on your network in each of the Windows NT local databases. If your network
had ten users and ten Windows NT computers, if you added one
new computer, you would have to recreate all of your ten users within that new computer’s local user account database. Likewise, if a
new user is added to the network, their name would have to be
added to all of the local databases on each of the existing Windows
NT computers. The same would go if a user changed their password;
each Windows NT computer would need to be updated with the
new password of the user.
The workgroup networking model is highly decentralized, and
requires an administrator to perform many repetitive tasks, such as
351
16.919-x CH10.i
352
8/28/98 1:47 PM
C h a p te r 1 0
Page 352
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
adding user accounts or having users create many different shares
and assigning passwords to these shares. The workgroup model for
Windows NT is similar to the Windows 95 workgroup model, in
that it is often only found on very small networks.
Bindery-Based Model
The bindery-based model is one that is used by Novell NetWare versions up to NetWare 3.2 (all Novell servers that are version 4 or
higher use Directory Services). Bindery-based networks follow the
client/server model of networking. Novell bindery-based servers still
have a large presence in many networks to this day.
In a bindery model, there is one server and many clients. The server
contains a flat user account database (see Figure 10.2). A flat user
account database is one that contains the names of users, in one single list from A to Z, who are allowed to log onto the system. Also,
this database of user accounts is used to assign who has rights or
privileges to use different resources on the network. These rights are
either assigned on a user-by-user basis or a group-by-group basis.
The server is also responsible for containing all of the services on
the network. The client machines are not designed to provide any
FIGURE 10.2
Server
A Bindery-based network has a centralized user
account database. Client machines run no
services.
User Account Database
Name
Bob
Mary
Colleen
...
Password
Pass
cow
Bodybuilding
...
Logon to server
Name: Mary
Password: cow
Client
Client
Client
16.919-x CH10.i
8/28/98 1:47 PM
Page 353
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
services at all. This allows for a more centralized method of management of the network.
A client machine on this system is one that has a redirector installed
on it, such that it will connect to a central server, and try to authenticate against that server’s user account database. The user will supply a valid name that exists within the user account database (logon
name) and an associated password.
If the name and password exist within the server’s user account database, the user is granted permission to use the network, and in turn,
the user’s computer is given a “key” by the authenticating server.
This key is similar to a security badge that you may wear when touring a secured facility. The key essentially identifies who you are as an
individual and what groups you belong to. Based upon this key, you
as a user will be granted access to shared resources or denied shared
resources. This granting of the key is transparent to the user; they
will in most cases “see” those resources available to them and not see
the resources that are not available to them.
In a bindery model, there is a benefit of having one centralized database from which to perform all of the management tasks. This is
preferred in a larger network for managing resources, as it simplifies
and centralizes the management of these resources. There is also the
benefit of having the capability to give access to shared resources on
a user-by-user or group-by-group basis as security can be done on a
more specific basis. Also, users will not have to remember a whole
host of passwords to access different resources, but instead get seamless access or denial of resources that they wish to use.
A problem with the bindery model is when you have many servers
on the network. Bindery models do not allow for the sharing of
database lists between servers, as each server maintains its own user
account database. Because of this limitation, as more servers are
added to the network, every time a new user account is created, it
would have to be added to the user account database on each server
that would contain a resource being shared out to that user. Thus, as
a network grows, an administrator would need to do repetitive tasks
in order to maintain the security of the network.
A second issue with a bindery-based network system is that this
administrative model uses different utilities to perform different
functions. Thus, one utility is used to manage the file system, another to manage the users, and even another to manage the printers.
353
16.919-x CH10.i
354
8/28/98 1:47 PM
C h a p te r 1 0
Page 354
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
Every time a new service or resource is added, a new administrative
utility often must be learned to manage that resource or service.
NOTE
Domain Model
A Focus on Windows NT Although
other network operating systems
besides Windows NT use the Domain
model, this section will focus on
Windows NT for discussion purposes.
The domain model is another client/server model that is used in
Windows NT Server and OS/2 networks. It is similar to the bindery
security model, in its centralized administration of user accounts and
flat list of user accounts, but scales better for larger networks.
The domain model is a security model that uses a flat user account
database similar to the bindery model. The main difference is that
this database is stored on one or more computers known as domain
controllers (see Figure 10.3).
When a Windows NT server is installed, one of the parameters that
must be configured is what role the server should assume. There are
three possibilities:
á Primary domain controller (PDC)
FIGURE 10.3
á Backup domain controller (BDC)
A domain contains Domain Controllers that
store the user account database.
á Member server
Domain: Domain1
User Account Database
Name
Password
Tom
Pass
Colleen
bobo
...
...
PDC
BDC
BDC
Name: Tom
Password: Pass
Logons
Client PCs
Name: Colleen
Password: bobo
16.919-x CH10.i
8/28/98 1:47 PM
Page 355
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
Primary and backup domain controllers perform essentially the same
function. It is their role to store the user account database. The difference is that a PDC stores the master copy of this database. It is in
this master copy that changes can occur. If a new user was added,
the PDC’s database would be affected. The backup domain controller’s user account database is a replicated copy of that from the
PDC.
There can only exist one PDC in a domain, yet you can specify as
many BDCs as you wish. At any time a BDC can be promoted to a
PDC, and thus a PDC demoted to a BDC. Issues that involve the
number of BDC’s, the placement of the domain controllers, and so
on are covered in books and courses relating to the Windows NT
Server and the Windows NT Enterprise exams.
The role of a member server is that it contains resources such as files,
printers, and applications that users may wish to access on the network. It in itself does not store a domain user account database, but
instead gives access to resources to users based upon the users drawn
from the list of user accounts on the domain controllers. The member servers are also not involved with processing logon request for
client machines, as this is a function that is only performed by
domain controllers.
By placing the shared resources out on the network on member
servers, you are placing these resources on computers that are not
allocating their system resources such as RAM, CPU power, and disk
space to process and maintain the user account database and the
user’s logon requests.
Some drawbacks to the domain model are that there is a separate
utility to perform different administrative functions, depending on
the resource being administered. That is, similar to the bindery
model, there is one utility to create users in and another to manage
printers with. If, for example, you were to install a fax server, there
would be a new utility to manage this fax server. Another important
issue with the domain model is that a domain is not designed scale
to more than 40,000 accounts. An account is every registered user,
group, and computer within the domain. Thus in some larger networks, multiple domains would need to be created.
355
16.919-x CH10.i
NOTE
356
8/28/98 1:47 PM
C h a p te r 1 0
Page 356
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
NTDS versus x.500 Directory
Services Microsoft has recently
begun calling its domain model
“Windows NT Directory Services” or
“NTDS.” This is simply a naming convention. The domain model, or NTDS,
does not provide the features and
functionality of true x.500 Directory
Services. This will be part of Microsoft
Windows NT 5.0 Active Directory
Services. For more information on
Directory Services, see the following
section.
Directory Services Model
Directory Services, also known as the X.500 standard, is the latest in
security management to be offered for networking security. It currently is used by Banyan Vines, Novell NetWare 4.x and higher, and
is to be incorporated into the release of Windows NT 5.
Directory Services is a powerful security management system for a
network, as it can accommodate a small to extremely large network.
It solves many of the limitations found in the workgroup, bindery,
and domain security models.
Directory Services is based upon a hierarchical distributed database
model (see Figure 10.4). This model allows for the management of
all resources through one utility, as well as providing a high level of
fault tolerance within the system.
Management on a Directory Service security system is based on a
hierarchical user account database. The idea behind this is similar to
a file system database. No one stores all of their files in one directory
on their hard drive. Files are grouped together into directories such
that files that go together or are related to one another are placed
together for management purposes and ease of reference. The same
Company
Servers
NA
Asia
LA
NY
Bob
L.A.
Europe
Pete
Mary
Printer1
FS1
Tokyo
FIGURE 10.4
Directory Services has a distributed hierarchical
database.
16.919-x CH10.i
8/28/98 1:47 PM
Page 357
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
can be said for a Directory Service user account database. Instead of
directories, containers store users together that work together or
access the same resources together. In fact, many Directory Services
databases are organized in a manner that is similar to their corporate
organizational charts.
The management of resources is not limited to users and groups
within a Directory Services database. Other resources on the network also have objects within the database. Thus when a printer is
installed, its object is placed into the database as well. Management
of this printer can also be done from the Directory Services database.
So could the management of a fax server or any other device on the
network. This functionality allows administrators to use one utility
to do most of their network management.
The third main benefit of Directory Services is that it allows the partitioning of the database such that portions of it, partitioned around
the containers, can be placed on different servers. This would mean
that if a user is added in Los Angeles, the server in London, England
does not need to be updated. This feature allows for the minimization of network traffic over slow WAN links. In a domain model, all
user accounts are copied to all BDCs whenever a user account is
added.
In short, the Directory Services model is the standard that all operating systems are migrating toward, as this model has features that can
ease administration of the network, as well as allow a network to
scale to any size desired.
In summary, here are the following features of the four administrative models:
á There are two derivatives of workgroup models—one used by
Windows 95, the other by Windows NT.
á Windows 95 workgroup models have no user account databas-
es; all security is done with the use of passwords.
á Windows NT workgroups have a decentralized user account
database requiring administrators to perform repetitive administrative tasks such as creating users multiple times.
á Bindery-based systems use a flat user account database. This
user account database is not shared with other servers. You
must use different utilities to manage different resources.
R E V I E W
B R E A K
357
16.919-x CH10.i
358
8/28/98 1:47 PM
C h a p te r 1 0
Page 358
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
á Domain-based systems are similar to bindery-based systems,
with the exception that more than one server shares the same
user account database.
á Directory Services allows for a distributed hierarchical database
EXAM
TIP
shared by all servers, from which all resources can be managed.
Conceptualizing User Accounts,
Global Groups, and Local Groups
Throughout your MCSE testing,
there are a few very basic conceptual frameworks you must understand. One of the most important
is the interaction of user accounts,
global groups, and local groups.
This chapter defined users and
groups earlier, but you must look
more deeply into the actual functional value of each.
MANAGING USER ACCOUNTS AND
GROUPS USING WINDOWS NT
As seen earlier in this chapter, when looking at the Networking security models, the Windows NT domain model is designed to provide
far greater security than Windows 95. As such, Windows NT, using
the domain administrative model, is the centerpiece of any Microsoft
network where security is a major issue. An organization might not
feel that its everyday documents require security, but most companies have payroll information or other data that they want to guard
from access by unauthorized individuals. This section will focus on
the Windows NT domain administrative model in describing how a
domain is managed and how security for files and printers is accomplished.
User Accounts
In most instances, a user account is created for each individual on
the network and is meant for use only by that one person. This is
done through the “User Manager for Domains” utility (as seen in
Figure 10.6 in Exercise 10.1). This account generally is a contracted
form of the person’s name or some other unique value, and no two
users can have the same username in a single user account database.
At their most basic level, a user account usually contains values for
the following three properties:
á A username. This element distinguishes one account from
another. This property requires a value.
á A password. This element confirms the user’s identity. Indi-
vidual passwords should be kept private to avoid unauthorized
16.919-x CH10.i
8/28/98 1:47 PM
Page 359
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
access. This property may be optional, depending upon security restrictions.
á The groups of which the user is a member. These groups
determine the user’s rights and permissions on the network.
This is an optional property.
A number of other optional properties, such as a home directory (a
place where a user can store personal files on the network) or specific
information about the user such as their full name or description,
also exist. None of these properties are crucial to the functioning of
the account in the way that the elements enumerated above are. In
Exercise 10.1, you create a very basic user account and observe some
of the available options.
D E FA U LT AC C O U N TS I N WI NDOWS NT
When Windows NT is first installed, there are only two user
accounts that are created. The first one is named Administrator,
and you are prompted to supply a password for this account during
the installation process. The second account is called Guest. This
account is disabled by default.
There can be a third account created, depending on whether or not
you install Internet Information Server (IIS) during the Windows NT
install. The name of this account is IUSER_Computername, where
Computername is the NetBIOS name of the computer that IIS is
installed upon.
In the creation of user accounts and their passwords, you must strike
a balance between security and user friendliness. Passwords have settings such as expiration dates, uniqueness, and how often they must
be changed. Setting these options such that a password must be
changed on a basis that is that is too frequent or one that requires
too many long, unique passwords is almost certain to result in a less,
rather than more, secure environment. If users are unable to remember such a password, they often simply will stick a note to their
monitor with their password on it or come up with some other
highly insecure way of jogging their memory. If this starts happening, you know that your policies are probably too stringent.
359
16.919-x CH10.i
360
8/28/98 1:47 PM
C h a p te r 1 0
Page 360
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
Groups
Now that a user has been established, the next step in granting that
individual access to resources is to assign proper permissions. To ease
the management of all users in the system, this should be done by
creating a group or a set of groups, assigning permissions to the
groups, and then placing the user inside the appropriate groups. As
mentioned before, it is easier to manage five groups than five hundred users.
By default, Windows NT creates a number of built-in groups that
are defined with the rights necessary to perform particular tasks.
These groups are task-specific and are inherently different from the
type of groups you normally create, which are resource-specific.
These kinds of groups are discussed in more detail in the next Indepth.
Windows networks can include two types of the resource specific
groups: global and local. Each of these has very specific functions.
NOTE
Global
Local Users in Windows NT
Windows NT does give you the ability
to create local user accounts that
cannot be added to a global group.
For more information on this and
other Windows NT issues, consult a
book that focuses on Windows NT
Server, like those mentioned in the
“Suggested Readings and Resources”
section of this book.
Global groups, like user accounts, are created only on the primary
domain controller of a Microsoft domain. Backup domain controllers receive a copy of this database; thus, they also contain global
groups. These groups function primarily as containers for user
accounts. Global groups are designed to contain general groupings of
people, such as Sales, Accounting, or the IS department. Global
groups cannot contain other groups—only users from the domain in
which the global group is created are permitted to be part of a global
group.
Local
Local groups, on the other hand, can be created on Windows NT
Server or Workstation and can include both user accounts and global
groups. Moreover, these groups are assigned permissions (see next
section for more on permissions).
The premise behind local groups is that an administrator—who is in
charge of a server, such as the accounting server—can create a local
group for this server’s specific management needs. An example might
be that the local server has two directories shared out. One directory
is called Bonuses and the other is called Corporate Policies. This
16.919-x CH10.i
8/28/98 1:47 PM
Page 361
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
administrator could create two local groups, one called Restricted
and one called General.
For the local group Restricted, this group may contain only the
global groups of Sales Managers, Accounting Managers, and Human
Resource Managers. The local administrator then would assign permissions to the Bonuses directory by using this local group called
Restricted.
For the local group General, this group may contain all global
groups, and this local group called General may be given the permissions to the Corporate Policies directory.
In Exercise 10.2, you create both types of groups and explore how
they interact with users and resources. Note that this exercise
assumes you are using a Windows NT domain controller. If this is
not the case, you will be unable to complete the steps as written. In
that case, you can participate in the creation of the local group and
ignore instructions that deal with global groups.
B U I LT- I N G L O B A L A N D L OCA L GROUP S I N
W I N D OW S NT
Windows NT also contains some built-in global and local groups.
These groups are given permissions and rights to various components on the network in order to provide some general functionality
on the system. Users can be added to or removed from these
groups.
These global groups only exist on domain controllers.
á Domain Users. All users created within a domain are placed
into this group.
á Domain Admins. The Administrator account is placed within
this group. All domain-wide administrators should also be
placed in this group.
These local groups are found on domain controllers, member
servers, and Windows NT workstations that are part of a domain.
á Administrators. Contains the Domain Admin global group.
This group can manage all security and resources on the
computer.
á Users. This group contains the global Users group.
continues
361
16.919-x CH10.i
362
8/28/98 1:47 PM
C h a p te r 1 0
Page 362
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
continued
á Guest. This group contains the guest account.
á Account Operators. Members of this group have the ability to
create users and groups, both global and local. This group
cannot manipulate the Administrator, the Administrators
group, the Domain Admins group, or the Server operators
group. This group is only found on domain controllers.
á Backup Operators. Members of this group have the rights
needed to backup and restore files on the computer.
á Print Operators. Members of this group can manage all printers on the computer. This group is only found on domain controllers.
á Server Operators. Members of this group can share and stop
sharing resources on the server, backup and restore files on
the server, and shut down the server. This group is only found
on domain controllers.
á Replicator. This group is used with the Directory Replicator
Service.
Usually, the default rights associated with these built-in groups will
be fine to perform the functions for which they are intended. The
Administrators, Server Operators, Backup Operators, Print
Operators, and Account Operators groups all have the right to log
on to Windows NT Server interactively. This feature is not granted
to other users by default on Windows NT Server.
For managing resources, you create the group and add users to it,
at which time the group is ready to be given permissions in the file
system, such as Read permissions to a directory or Print permissions to a printer.
Windows NT also creates four special groups, each of which has
special uses and access privileges. You cannot delete or rename
these groups. You cannot add or remove users from these groups.
Because of this, these groups do not appear in User Manager for
Domains. Based upon where you are or what you are doing, the
system will associate you with one or more of these groups. You
can give or deny these special groups permissions to resources.
The following list details these four groups:
á Everyone. This umbrella group includes all users on the network.
á Creator-owner. If a user creates or owns a directory, he gains
whatever rights are given to this group, as each file and directory is always associated with an owner.
16.919-x CH10.i
8/28/98 1:47 PM
Page 363
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
á Interactive. This group is fluid, in that a user becomes a part
of it when they access a local resource, and they are excluded from it when accessing a resource over a network connection.
á Network. This group is exactly the opposite of an interactive
user group. This is another fluid group that includes any user
that is accessing a resource over a network connection.
Creating groups and users provides the base upon which the rest of
your security is built. You should now know what a user is, and how
users and groups interact. Do not get overly caught up on the different groups and their abilities. That is the purpose of the Windows
NT Server exam. The next section explores using these groups and
users to give or restrict access to network resources.
Permissions
Permissions refers specifically to the level of trust that the owner of a
resource has in the people with which he or she shares the resource.
Although very subtle permissions structures can be constructed
using Windows NT and Windows 95, a resource generally will
either be shared as read-only or full-control. By default, both
Windows NT and Windows 95 share resources with full control,
which means that others cannot only view your shared resources but
also can append, modify, and even delete them. For the less-trusting
owner, a good compromise is to grant read-only permissions, which
enable others to view your files or print to your printer, but not to
modify those files or change the printer’s settings.
Rights
The difference between having rights and receiving permissions
might seem like nothing more than a matter of semantics, but this is
not the case. In Microsoft terminology, rights are general attributes
that particular users or groups have. These rights include the capability to log on locally or to load and unload device drivers. These
particular user rights make administrators more powerful than users.
Permissions refer to the level of control a particular user or group
has over a specific resource. Examples of resource control would be
363
16.919-x CH10.i
364
8/28/98 1:47 PM
C h a p te r 1 0
Page 364
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
the “Read” permission to a file or directory, or the ability to “Print”
to a shared printer.
MORE ON RI GHTS
If a user and an administrator have full-control access to a directory, either of them can read, modify, or even delete that resource. If
the directory must be restored from tape, however, only a member
of the Administrators, Server Operators, or Backup Operators
groups can accomplish this task. By default, only these groups
have the right to restore files and directories. To see the different
rights available to Windows NT users, select the Policies menu in
User Manager for Domains and then select User Rights. Choose
the check box in the lower left to view additional Advanced Rights.
All these terms might make network security seem a bit daunting,
but this is not necessarily so. Perhaps it is easiest to think of server or
workstation resources just as you would think of anything else that
you must care for and protect.
For instance, imagine that you have a house. If you want, you can
just keep the house to yourself and not admit entrance to anyone
else, thus preventing damage to your possessions. Of course, you also
can allow others to enter, but then you take the chance that someone
might damage your possessions, either maliciously or inadvertently.
Because of this, it’s a good idea to take some precautions about who
you invite to your house. Moreover, you almost certainly will be
more watchful of some guests than others, and you will seek to protect certain rooms or possessions more than others. Lastly, because
you can’t watch everyone all the time, you probably will want to
have some good locks on the doors and sufficient insurance against
theft or disaster.
IMPLEMENTING SECURITY
WINDOWS NT
ON
Compare user-level security with access permission assigned to a
shared directory on a server.
16.919-x CH10.i
8/28/98 1:47 PM
Page 365
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
The previous section discussed security issues on a Windows NT
network; this section will look at how to implement security on a
Windows NT network. The main areas of focus will be on how to
implement security on shared directories, on NTFS partitions, and
on printers being shared by a Windows NT computer.
Creating and Assigning Permissions to
a Shared Folder in Windows NT
In Exercise 10.3, you create and share a directory called Public. The
group Everyone is given Read permissions to the directory, and the
group Local Training is given Full Control. Remember that only
directories can be shared, and all files and subdirectories within that
directory are available over the network through the share. Exercise
10.3 assumes a FAT partition with no NTFS file-level security, or an
NTFS partition on which no restrictions have been set. Remember
that NTFS is the native Windows NT file system, and that it allows
for additional security beyond what the FAT file system can offer.
G E N E R A L A D M I N I S TRATI V E RULES
Rights and permissions also can be given directly to user accounts
themselves, but this is not recommended. Not only is such security
cumbersome, but it is also difficult to administrate and troubleshoot.
If you have a user who has specific resources that are different
than those of anyone else on the network, resist the urge to simply
assign that user the needed permissions directly. Rather, create a
new set of groups (local and, if needed, global), place the user into
the proper group or groups, and then assign permissions and rights
through the local groups as needed.
This might seem unnecessarily redundant, but it can be very useful
later on, especially if the user leaves your organization and is
replaced by a new user who now needs the same permissions. The
new user simply can be placed into the needed group(s), and the
old account can be removed from them. If, on the other hand, you
have assigned file system permissions directly, you then must hunt
down all the directories to which the original user had access,
remove the old user from each one, and then insert the new user.
365
16.919-x CH10.i
366
8/28/98 1:47 PM
C h a p te r 1 0
Page 366
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
Assigning File-Level Permissions on an
NTFS Partition
If you are using the standard FAT file system native to DOS,
Windows, and Windows 95, your Windows NT security structure
will be complete after you assign share-level permissions to your files.
In Exercise 10.4, however, assume that the partition on which the
share is located is formatted with NTFS, Windows NT’s native file
system. In this case, you can assign additional rights within the share
on a per-directory and even per-file basis. The strength of NTFS
security is two-fold:
á NTFS security gives the administrator a wider range of flexibil-
ity in assigning rights to files and directories.
á NTFS security provides security even at the local level, some-
WA R N I N G
thing that a FAT partition does not support. Interactive users
are unaffected by share-level security options, but still are limited by NTFS file-level security.
No Access The exception to the
principle of additive privilege is the
No Access permission, which immediately blocks all other rights.
Because of this, the No Access
option should be used sparingly
and very carefully. Numerous No
Access permissions on the network
usually point to a poorly implemented security structure. If you don’t
want users to access a resource, it
is sufficient simply to not give them
permission—explicitly banning the
users access generally is overkill.
Also, never implement No Access
for the Everyone group—this group
includes you, as well as all other
administrators and users, none of
whom will be able to get to the
resource until the No Access is
removed, even if they belong to
other groups that do have sufficient
permissions.
In the Public folder shared in Exercise 10.3, you see that two sharelevel permissions exist for this directory:
á Everyone: Read
á Administrators: Full Control
In Exercise 10.4, you will assign a new permission to this directory,
this time through NTFS security. The permission to be assigned
will be:
á Everyone: Change
You should always consider how this change will affect the permissions of the Everyone and Administrators groups before altering your
permissions structure. Remember that Read permissions allow Read
(R) and Execute (X) permissions, while Change grants these permissions plus Write (W) and Delete (D). Likewise, Full control offers
these permissions plus Take Ownership (O) and Change Permissions
(P). Share-level rights and file-level rights are both cumulative within
themselves. For instance, an administrator on a Windows NT network will be a member of both Administrators and Everyone—and
possibly a number of other groups as well.
16.919-x CH10.i
8/28/98 1:47 PM
Page 367
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
367
In your Public share, the user would gain RX from the Everyone
group and RXWDOP from the Administrator group. The user then
would have RXWDOP over the share. On the other hand, if you
include the NTFS permissions for Everyone, the user has RXWDOP over the share and only RXWD at the file level. Under NTFS,
only permissions granted at both the share level and the file level will
be applied, and the administrative user will have only Change
(RXWD) permissions over the share.
Generally, you will try to use a Windows NT domain model to provide resource access on your network. In some situations, though,
you may need to implement a workgroup sharing model or use a
Windows 95 machine as a server. The main reasons for this is
because the size of the network is not large enough or the security of
the data is not critical enough to justify the expense of Windows NT.
Printer Sharing with Windows NT
The second major resource with which you will be expected to be
familiar is the printer. For many administrators, printers have been a
constant trouble spot, and it seems that a disproportionate percentage of network problems are caused by these devices. Because many
of these problems are due to improper or modified printer configurations, careful setup and effective security structures can save an
administrator considerable time in this area.
Windows NT has the capability to enable remote users to dynamically download print drivers specific to their own system into RAM
each time the users print. This allows for easy driver updates and
also enables users to connect to a new printer without having rights
to install drivers on their local system. This process is called connecting to a printer.
As you install and assign permissions to the printer in Exercise 10.7,
observe that many of the processes are very similar to the steps you
took in creating, sharing, and securing files and folders. The groups
and users that you can draw from are the same in granting Printer
permissions as they are when granting File permissions.
NOTE
To connect to a network printer, you first must install and configure
the printer on a server. Every network printer, in other words, is just
a local printer that has been shared by its owner.
Drivers Remember that a driver is a
piece of software that allows a piece
of hardware, in this case a printer, to
communicate to the operating system, in this case Windows NT or
Windows 95.
16.919-x CH10.i
368
8/28/98 1:47 PM
C h a p te r 1 0
Page 368
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
IMPLEMENTING SECURITY
WINDOWS 95
ON
As noted earlier, Windows 95 also can act as a server, albeit in a less
robust capacity. Windows 95, under the workgroup model, was
shown to have one type of security. Windows 95 can actually support two types of security: share-level and user-level. Share-level
security is supported in a workgroup model or when a Windows 95
machine is part of a Windows NT domain. User-level security is
only supported when Windows 95 is part of a Windows NT
domain, or in the presence of a Windows NT server or workstation,
or as a client in a Novell bindery-based security model.
As you read about the different security models, notice that
Windows 95’s user-level security is nearly identical to Windows
NT’s share-level authentication. Moreover, notice that Windows 95
does not support file-level local security as provided by Windows
NT NTFS partitions, nor does Windows NT have the ability to
provide Windows 95’s low-security, password-only share-level
option.
Share-Level Security on Windows 95
Under Windows 95’s simple share-level security, passwords are
assigned to permit access to each directory or printer share. To access
the share, a user must supply the correct password.
When creating a shared directory using share-level security, you can
grant one of three types of access:
NOTE
á Read-only access. After entering the correct password, a remote
No Password If no password is
entered, all users have full or readonly access to the directory, depending on which option was specified
when the shared directory was
created.
user can access a directory, its subdirectories, and its files.
However, the user cannot delete files or write files to that
directory.
á Full access. A remote user who supplies the correct password
has read and write privileges to that directory and all its files
and subdirectories.
á Depends on password. Two different passwords can be created:
one allowing read-only access, and one allowing full access.
16.919-x CH10.i
8/28/98 1:47 PM
Page 369
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
The type of access granted to a user depends on the password
that the user has supplied.
In Exercise 10.5, you create a directory share using share-level security. Remember that a share is an entry point on your computer
from which you can give others access to your local resources.
Print queues also can be shared with other network users using
share-level security. If a password is specified for the share, a network
user must enter that password to access the print queue and connect
to that printer. If a printer is shared with a blank password field—
meaning no password was entered—any user can connect to and
print to that printer.
Because share-level security relies on access passwords, this form of
security has the following disadvantages:
á To access different shares, a network user must know numer-
ous passwords.
á Passwords can easily be forgotten. Windows 95 can cache pass-
words so a user does not need to enter them each time.
However, if the creator of the share forgets the password, then
the password must be changed to enable another user to access
the share.
á Nothing prevents a user from disclosing the password to an
unauthorized user.
User-Level Security on Windows 95
User-level security can be used to overcome the shortcomings of
share-level security, and where it is available, this type of security is
generally the preferred security structure. With user-level security,
you can grant specific user accounts or group accounts to a shared
directory or printer. Instead of relying on a password that could be
used by anyone, the user account accessing a shared resource must
be authenticated to ensure that that account has been granted access.
User-level security, therefore, provides a level of personal flexibility
and accountability that is not available with share-level security.
Windows 95 cannot manage user accounts by itself as it does not
have the ability to house and manage a user account database.
369
16.919-x CH10.i
370
8/28/98 1:47 PM
C h a p te r 1 0
Page 370
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
Instead, the application must use another authentication database,
such as that of a Windows NT or a NetWare server, which can
authenticate the user trying to access the resource. In user-level security, Windows 95 must defer to a machine with a user database and
present all requests for access to that machine for authentication.
US E R-LE V EL S E CURI TY
Windows 95 ships with the ability to perform user-level security with
Novell NetWare and Windows NT. Other developers can write their
own services to perform user-level security as well.
To initiate user-level security, the Windows 95 computer must
obtain a copy of the accounts list from one of the following sources:
á Windows NT Server 3.5 (or later) computer
á Windows NT Workstation 3.5 (or later) computer
á Windows NT 3.5 (or later) domain controller
á NetWare 3.x server
á NetWare 4.x server with bindery emulation enabled
When a directory is shared with user-level security, the users or
groups to be granted access to the share are assigned privileges. You
can grant each user or group one of the following privileges:
á Read-only. Users can access files and subdirectories in a directo-
ry, but cannot delete or save files to that share.
á Full access. Users can read, write, and delete files in the directory.
á Custom. Any number of the following privileges can be granted:
• Read Files
• Write to Files
• Create Files
• List Files
• Delete Files
• Change File Attributes
• Change Permissions
16.919-x CH10.i
8/28/98 1:47 PM
Page 371
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
Exercise 10.6 demonstrates how to grant a network user access to a
directory share. For this exercise, you must be part of a domain that
contains a server with a user accounts database or a stand-alone
Windows NT Server or Workstation account database that could
be used. If the user accounts exist on a NetWare server, you will
need to install the Client for NetWare Networks, the IPX/SPXcompatible protocol, and File and Printer Sharing for NetWare
Networks, and make your selections accordingly throughout the
exercise.
You now have learned what a user and a group are, and how they
can be used to provide network access and file security. You have
seen the way that both Windows 95 and Windows NT handle security issues, and should be able to see some of their major differences.
Remember that the same principles that guide file sharing also work
for the other major network resource printers that you will examine.
Printer Sharing with Windows 95
To use Windows 95 as a network print server, the printer must be
attached to the Windows 95 machine locally and configured with
the proper driver, just as it would be if it were serving only local
users. The printer then must be shared to enable other users to
access it. To share a printer in Windows 95, a 32-bit, protectedmode client, and the “file and printer sharing” service must be
enabled.
Exercise 10.8 demonstrates how to share a network printer from a
Windows 95 machine. This exercise assumes that you already have
installed and configured a printer.
When the printer has been configured and shared on the network
print server, a Windows 95 client can be configured to connect to
the print server and print to the printer over the network. This configuration can be established either manually with the Add Printer
Wizard or by configuring the network printer for Point and Print
setup.
Printer security can support either share-level or user-level security,
depending on the security role the computer is part of, as described
in the previous sections. Share-level security requires that the user of
a printer be able to provide a password in order to print to the
shared printer. User-level security will allow the administrator to
371
16.919-x CH10.i
372
8/28/98 1:47 PM
C h a p te r 1 0
Page 372
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
select, on a user-by-user or group-by-group basis, who can or cannot
print to the shared printer.
ADDITIONAL ADMINISTRATIVE TASKS
Besides setting up the network and making sure that your users have
access to what they need (and can’t get to things they don’t), an
administrator also has a number of other important day-to-day tasks
to fulfill. The remainder of this chapter gives you a brief introduction to the following responsibilities:
á Auditing
á Handling data encryption
á Handling virus protection
á Securing equipment
Auditing
Another option you might need to consider is auditing, which is the
process of creating a database that records particular events that
occur on your network. Generally, you can decide what events to
audit, from application information to security options. Figure 10.1
shows one of many different auditing windows in Windows NT.
The utilities to perform auditing come with Windows NT and
Windows 95. These tools are known as Event Viewer and
Performance Monitor in NT, and both are discussed in greater detail
within Chapter 11, “Monitoring the Network.” Windows 95 does
come with a utility called System Monitor. There are also a variety of
third-party auditing tools available on the market.
Handling Data Encryption
Usually, the file and share security discussed previously is more than
adequate. However, if your network is used for especially sensitive
data and you want to prevent anyone from stealing information, you
FIGURE 10.5
A directory auditing window in Windows NT.
16.919-x CH10.i
8/28/98 1:47 PM
Page 373
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
Handling Virus Protection
Much like humans, computers are susceptible to certain types of
viruses. Unlike those that strike us, though, computer viruses are
created intentionally with the aim of injuring or altering your
machines. Viruses can be spread through computer systems in many
ways, but the most common is through an executable file. Having a
good virus scanning program—none come with any Microsoft program—is a necessity for an administrator. Numerous third-party
companies make virus-scanning software, including Norton and
MacAfee, to name just two.
Securing Equipment
You might think that if you have taken care of backup, RAID,
shares, NTFS permissions, virus scanning, and encryption, your
data is completely safe. There is, however, one more thing of which
you should be sure. Any computer is far more insecure if people can
get to its server, so you always should lock your server in a room
that only authorized personnel have access to. Having the server out
in the open provides a security risk, such that it is open to anyone to
tamper with it. Most companies have a “server room”—often a large
wiring closet—where all server machines are stored. Make sure this
location is neither too cold nor too hot, that it has adequate ventilation, and that only authorized individuals have access to it.
Additionally, whenever you make a change to the network, be certain to document the changes you have made. This can make troubleshooting and maintenance far easier and can save you valuable
time. See Chapter 11 for more information.
NOTE
can take an additional security measure by forcing data encryption.
Encryption codes the information sent on the network using a special algorithm and then decodes it on the other end. This technique
offers varying degrees of safety, largely based on the length and complexity of the code used to encrypt the data. With the advent of the
Internet, encryption technology is becoming more important and is
key in the new “secure transactions” toward which companies on the
World Wide Web are now working.
Anti-Virus Software Anti-virus software cannot simply detect any virus;
rather, this software generally is
designed to look for particular infections. Because of this, scanning
software is updated regularly, often
at no extra charge. Even if it does
cost a bit, though, always keep your
virus-checking software as new as
possible.
373
16.919-x CH10.i
374
8/28/98 1:47 PM
C h a p te r 1 0
Page 374
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
C A S E S T U DY : W H I C H M O D E L
FOR
SECURITY?
ESSENCE OF THE CASE
A N A LY S I S
The essence of the case is as follows:
This analysis will be presented by comparing a
Windows 95 workgroup to a Windows NT domain
model. It will conclude with a table contrasting
the two solutions.
• There will be eleven users sharing files
on the system.
• A Windows 95 workgroup and Windows
NT domain model are to be compared.
• The company plans on expanding in the
future.
• There is some very confidential information within the law firm.
SCENARIO
You have been called in to a small law office. This
law office has five lawyers, five legal secretaries,
and a receptionist. The law office has done everything the old-fashioned way; that is, using the old
typewriter, pens, and paper. They feel that they
will be more productive if they could put all of
their documents on the network, thus allowing
them to find old cases faster, as well as letting
them produce legal documentation faster.
Windows 95 Workgroup
A Windows 95 workgroup solution would definitely be a cheaper option in terms of software
costs. All computers would have Windows 95
installed on them. Each user would be in charge
of maintaining share level security themselves.
The Windows 95 workgroup solution would
require that each user be trained and become
proficient in the management of their computer,
in order to have an effective security system in
place. Also, passwords for each shared resource
would need to be maintained by each user.
As this network grows, the ability to manage all
the files on this type of distributed network would
come under pressure. As people are added to
the system, the number of passwords for each
shared resource would increase.
This law firm wants all to share files back and
forth, yet they do want to have some degree of
security in who can access which files. Some of
the documents that they hold contain very confidential information. It is a small firm, yet they are
intent on expanding in the future.
The ability to fully secure information would be in
jeopardy. Because Windows 95 only supports
share-level security and not file system security,
anyone can go up to any computer and interactively obtain any document they want.
The firm has decided to use a Microsoft networking solution, but are not sure of whether to use a
workgroup model based on Windows 95 or a
Windows NT domain model. Your job is to present the issues of choosing one system versus
another, so that the two senior partners can
decide which option they would like to pursue.
Windows NT Domain Model
The Windows NT domain model would cost more
than the workgroup model, as you would need to
purchase Windows 95 for each desktop, yet you
would still need to purchase a Windows NT
server.
16.919-x CH10.i
8/28/98 1:47 PM
Page 375
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
375
C A S E S T U DY : W
A DHDI CTHI TM
L EO DHEELR F
EO R S E C U R I T Y ?
One benefit of having a Windows NT server running a domain is that all security and user
accounts could be centrally managed. A policy on
password changes could be enforced, and this
model would allow for growth over time.
other users. Still, a secure room should be created in which to store the Windows NT server.
Windows NT also supports auditing, and has an
auditing program built into the system. This
would also allow for a degree of security, as it
would allow an administrator to be able to see
when and who is accessing what documents.
Because users are granted access on a user-byuser or group-by-group basis, passwords for
resource access would not be required. Instead,
each user would only need to remember one
password, the one to log in to the system.
To manage the system, only one (or perhaps
two) user would need to be trained as system
manager.
Since Windows NT provides file system security,
sensitive documents could be placed on the
server, without fear of unauthorized access from
Comparing the Two Systems
Table 10.1 compares both systems.
TABLE 10.1
WORKGROUP VERSUS DOMAIN-BASED NETWORKS
Feature
Workgroup
Domain
Cost of software
Lower
Higher
Cost of hardware
Lower
Higher
Number of people that would need training
More
Less
File system security
No
Yes
Share-level security
Yes
Yes
Requires a secure server room
No
Yes
Provides better security
No
Yes
Provide auditing
No
Yes
As usual, there is no right or wrong answer or
option here. If this law office feels that the security of the documents is not that great, they can
opt for the workgroup model and save some
money.
If this law firm feels that security of the data is
important, they would more than likely go for the
domain model supplied by Windows NT. They
could justify the higher cost of the domain model
by citing their security requirements.
16.919-x CH10.i
376
8/28/98 1:47 PM
C h a p te r 1 0
Page 376
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
CHAPTER SUMMARY
KEY TERMS
• Resource
• Sharing
• User
• Groups
• Security
• Zone
• Workgroup
• Bindery
• Domain
• Primary domain controller (PDC)
• Backup domain controller (BDC)
• Directory Services
• Global groups
• Local groups
• File-level security
• Share-level security
• Auditing
• Encryption
• Viruses
You now have learned to create users and groups and to configure
sharing and security for Microsoft resources using either Windows
NT or Windows 95. You also can connect to either of these
machines to gain access to their shared files and printers.
Furthermore, this chapter introduced you to a few optional security
measures available for sensitive data.
Knowing how to create network resources through sharing is crucial
not only for the Networking Essentials exam, but for an understanding of practical Microsoft networking as well. Make sure that you
can implement each of these structures and understand how they
work. Experiment with permissions and user rights and make sure
that the relationship between groups and users is clear.
16.919-x CH10.i
8/28/98 1:47 PM
Page 377
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
377
A P P LY Y O U R L E A R N I N G
Exercises
10.1 Creating a User Account in Windows NT
Objective: Create a new NT user account.
Estimated time: 10 minutes
1. Click Start, Programs, Administrative Tools.
Choose either User Manager (Windows NT
Workstation) or User Manager for Domains
(Windows NT Server).
2. User Manager opens (see Figure 10.6). If this is a
new install, only two users appear in the top window. As you might expect, Administrator is the
default administrative account for the machine,
and Guest is the default account for anonymous
access by users who do not have a username and
password of their own. The Guest account is disabled by default and must be manually enabled
before it is usable. If IIS is installed as well, a user
account with the name of IUSER_<computername> (<computername> is the NetBIOS name
of your computer) is also present.
3. Click Policies, Account to prompt the Account
Policy dialog box to appear (see Figure 10.7).
Observe that, by default, passwords must be
changed every 42 days. In addition, no restrictions are made as to password length or uniqueness. Account Lockout is turned off. Here, you
can set some of the default security information
for your network. If you are concerned that
someone might try to break into your network by
stealing or guessing a user’s password, these settings should be set to restrictive levels. Leave the
defaults as they are and click Close to return to
the User Manager.
4. Click User, New User. The New User dialog box
appears (see Figure 10.8).
FIGURE 10.7
The Windows NT Accounts Policy window enables you to set
password characteristics.
FIGURE 10.6
Account administration is done through the Windows NT
User Manager for Domains program.
16.919-x CH10.i
378
8/28/98 1:47 PM
C h a p te r 1 0
Page 378
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
A P P LY Y O U R L E A R N I N G
8. Examine the check boxes below the Confirm
Password field. By default, the User Must Change
Password at Next Logon field is checked. The
first time that a new user logs on, he is asked to
provide a new password. This enables you to set
an initial password but then transfer security over
to the user by having him define his own access
password.
FIGURE 10.8
The Windows NT New User dialog box enables you to record
information about a new user.
5. In the top field, type in a unique username (in
this case, TestUser) for the new account. This
name can be between 1 and 20 characters and
cannot include spaces or any of the following
characters:
“ / \ [ ] : ; | = , + * ? < >
6. Two text fields enable you to identify the user for
which the account is being created. The Full
Name field generally defines the person, and the
Description field defines the role they fill in the
organization. Fill both of these fields with the
values Test User and Training Department
Manager.
7. In the password field, you may enter any combination of 1 to 14 characters of your choice, with
the same exceptions that apply to the creation of
user accounts. Enter PASSWORD in both the
Password and Confirm Password fields.
(Remember that all passwords are case-sensitive,
so it matters whether you type PASSWORD or
password.)
9. The User Cannot Change Password option generally is used only for guest or multi-user accounts,
thereby keeping one guest from changing the
password and locking all other guest users out.
Leave this box unchecked.
10. The Password Never Expires option is intended
for system or guest accounts that require a static
password. As you will see, system policies can be
set by requiring occasional password changes by
network users. This setting overrides such policies. Leave this box unchecked.
11. The Account Disabled option enables you to disable an account temporarily while a user is on
vacation or after he no longer is allowed access to
the network. Generally, this option is preferable
to simply deleting the account, at least until it
has been determined that the user definitely will
not use the account again. Leave this box
unchecked.
12. The row of buttons at the bottom of the window
contains additional configuration options. You
use the Groups button in the next exercise, but
the other buttons contain options beyond the
scope of this book. Ignore them for now, but it
would be a good idea to return later to click on
each of them in turn and investigate the windows
they spawn. Close each without making modifications. This idea is good to follow in all exercises
because the key to mastering any Windows-based
16.919-x CH10.i
8/28/98 1:47 PM
Page 379
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
379
A P P LY Y O U R L E A R N I N G
product is to know where to click to find the
option you need. You should get used to exploring all the tabs and buttons available, but don’t
change anything unless you know what will happen.
13. Click the Add button. Notice that although all
the fields clear, the Add User window remains
open. This enables faster creation of multiple
users.
14. Click Close to return to User Manager. You now
see a third user, which is the TestUser account
you just created.
15. Click User, Exit.
10.2 Creating Groups on Windows NT
Objective: Create new global and local groups and
assign accounts to them.
This lab can only be completed if you are on a
Windows NT domain controller, and the primary
domain controller is online.
Estimated time: 10 minutes
1. Open User Manager for Domains. Observe the
groups in the bottom window pane. Some groups
have globe icons, such as the Domain Admins
group. Others, such as the Administrators group,
have a computer icon. As you might suspect,
Domain Admins is a global group, while
Administrators is a local group.
FIGURE 10.9
The New Global Group dialog box enables you to enter members and a description for a global group.
4. Note the two boxes at the bottom of the screen.
Administrator, Guest, and TestUser are displayed
in the Not Members box. Choose TestUser and
click the Add button. TestUser moves into the
Members box. It should be also noted that if any
users were selected at the time you created the
group, they will also be members of this group.
5. Click Close to return to User Manager for
Domains.
6. Click User and select New Local Group.
7. The New Local Group dialog box appears (see
Figure 10.10).
2. Click the User menu choice and choose New
Global Group. The New Global Group dialog
box appears (see Figure 10.9).
3. Type Global Training in the Group Name field.
In the Description field, type Training
Department Members.
FIGURE 10.10
The New Local Group dialog box enables you to add users
and global groups.
16.919-x CH10.i
380
8/28/98 1:48 PM
C h a p te r 1 0
Page 380
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
A P P LY Y O U R L E A R N I N G
8. Enter Local Training in the Group Name field
and leave the Description field blank.
9. Observe the members list box, which is empty.
Click on Add.
10. The Add Users and Groups dialog box appears.
Choose the Global Training group and click Add.
Note that you also could have added TestUser to
the group directly. In Windows NT
Workstation—which does not support the creation of global groups—this would have been
your only choice. Click Close to return to User
Manager for Domains. Click OK to return to the
New Local Group dialog box. The members list
now includes the Global Training group. Click
OK to return to User Manager for Domains.
11. Click User, Exit.
10.3 Sharing a Directory on a Windows NT
FAT Partition
Objective: Share a Windows NT directory and assign
share-level security to it.
Estimated time: 15 minutes
1. Click on Start, Programs. Then click on the
Windows NT Explorer icon to bring up the
Explorer window. Double-click on the icon representing your C: drive.
2. Select the root of the C: drive and then rightclick on it to call a context-sensitive menu.
3. Select New, Folder. A folder appears under C:,
and you are prompted to enter the name of the
folder. Type Public and press Enter.
4. Click on the new Public folder (in the left window). The folder is highlighted, and the right
window is now empty.
5. Right-click in the right window to make a
context-sensitive menu appear. Select New, Text
Document. Name the document My Shared Doc.
6. Select the Public folder again. Click File,
Properties (or use the quick menu and select
Sharing from there) to call the Properties dialog
box.
7. Click on the Sharing tab. Note that the directory
currently is not shared.
8. Click the Shared As option. The Share Name box
fills with “Public.” You can change or leave this
initial name. In this case, change the share.
Replace Public with My Share to illustrate the
difference between a directory name and a share
name.
9. Observe the Maximum Connections option. This
option enables you to control the number of concurrent users accessing the folder. Leave the
default, which enables unlimited concurrent connections to the share.
10. Click on the Permissions button to call the Access
Through Share Permissions dialog box (see Figure
10.11). Observe that, by default, Everyone has
Full Control over the new share.
11. Select the Everyone group and click the down
arrow in the Type of Access box. The following
four selections appear:
• No Access. A member of any group with this
permission is banned from the shared
resource.
• Read. Members can list, read, and execute
files, but cannot modify or delete them.
• Change. Members can read, list, execute, and
delete files, but are not able to change file permissions or assume ownership of the files.
16.919-x CH10.i
8/28/98 1:48 PM
Page 381
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
381
A P P LY Y O U R L E A R N I N G
10.4 Setting NTFS Permissions on a
Shared Folder
Objective: Add NTFS security to the Public share.
Estimated time: 15 minutes
FIGURE 10.11
The Access Through Share
Permissions dialog box enables you
to determine the type of access for a
particular group.
• Full Control. Members have complete control
of the resources, assuming that they have sufficient rights to match their permissions.
12. Click on Read in the Type of Access window.
Observe that the permissions level for Everyone
in the main window reflects the change.
1. Click Start, Programs. Select Windows NT
Explorer to open the Explorer window. Choose a
directory on an NTFS partition. If you do not
have an NTFS partition, you cannot complete
this lab.
2. Create a directory called TestNTFS and then
right-click on it. Select the Properties option
from the menu to open the TestNTFS Properties
window.
3. In the TestNTFS Properties window, click on the
Security tab and then click the Permissions button to open the Directory Permissions dialog box
(see Figure 10.12).
13. Click on Add to call the Add Users and Groups
window. Select the Local Training group and
click the Add button. The Local Training group
appears in the lower window. Click the Type of
Access down arrow and select Full Control.
14. Click the OK button and observe that Local
Training has been added to the list of groups
with permissions to the share.
15. Click OK to close the window and then click
OK on the Public Properties application.
16. In a few seconds, a hand appears under the
Public folder, indicating that the folder has been
shared.
17. Test the share by connecting to it from a
Windows 95 or a Windows NT client.
FIGURE 10.12
The Directory Permissions dialog box enables
you to update or replace permissions for a
group.
16.919-x CH10.i
382
8/28/98 1:48 PM
C h a p te r 1 0
Page 382
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
A P P LY Y O U R L E A R N I N G
4. Observe that the directory currently has its
default permissions list with Everyone—Full
Control as the only entry.
5. Select Everyone. Click the down arrow on the
Type of Access field and choose Read.
6. Take note of the check boxes near the top of the
window. The Replace Permissions on Files option
is checked, while the Replace Permissions on
Subdirectories option is cleared. If you have subdirectories and want the new access permissions
to filter down through them, you must check this
box. Because no subdirectories exist in this
instance, the point is currently moot, so leave the
defaults as they are.
7. If you need to enter additional groups into the
list, you can do so by using the Add button.
Click this button and observe the Add Users and
Groups window. Select the Administrators Local
group and then click on the down arrow next to
the Type of Access drop-down list and observe
the expanded choices. Permissions are broken
down to more specific levels, and Special File
Access and Special Directory Access enable you to
mix and match permissions to suit your needs. In
reality, you rarely will grant a group only the List
and Delete permissions, but you can if you need
to. If, for instance, a user needs to be able to
write to a directory, but should not be able to
view, read, or modify files in that directory, only
Write permission would be given to him. Give
Administrators Full Control permissions.
8. Click OK to return to the Directory Permissions
window. Then click OK to set the new permissions and return to Explorer.
9. Click File, Exit to close Explorer.
10. Share the TestNTFS directory with Everyone—
Full Control permissions and log on to the share
from a remote machine and observe the permissions available when you log on as an
Administrator as opposed to a TestUser. You
should be able to modify, create, and delete files
across the share if you are logged on as an
Administrator, but you should be able to only
read and execute while logged on as a TestUser.
10.5 Sharing a Directory Using Share-Level
Security
Objective: Share a Windows 95 directory using sharelevel security.
Estimated time: 10 minutes
1. From the Start menu, choose Settings, Control
Panel to display the Control Panel.
2. Double-click on the Network icon to display the
Network dialog box (see Figure 10.13).
3. Choose the Access Control tab and then choose
share-level access control.
4. Select the Configuration tab and choose File,
Print Sharing to display the File and Print
Sharing dialog box (see Figure 10.14).
5. Select both the I want to be able to give others
access to my files check box and the I want to be
able to allow others to print to my printer(s)
check box to enable others to access your printers
and files. Then choose OK to automatically
install File and Printer Sharing for Microsoft
Networks. You may be prompted for the location
of the source files. If so, type in the location and
click on OK.
16.919-x CH10.i
8/28/98 1:48 PM
Page 383
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
383
NOTE
A P P LY Y O U R L E A R N I N G
Share-Level versus User-Level
Security Share-level security is used
by default when File and Printer
Sharing for Microsoft Networks is
installed. The next exercise demonstrates how this can be changed.
Conversely, File and Printer Sharing for
NetWare Networks must use userlevel security. The share-level security
option is unavailable if File and Printer
Sharing for NetWare Networks is
installed.
9. Choose Sharing from the context-sensitive menu
to open the Sharing tab of the Properties application, as shown in Figure 10.15.
FIGURE 10.13
Use the Windows 95 Network Application to add and configure networking components.
10. Accept Password as the share name and choose
Access Type: Read-Only. Enter a password for
Read-only access of read and choose OK. The
FIGURE 10.14
Windows 95 File and Printer Sharing options offer you the
chance to grant access to your files and printer.
6. Choose OK and restart the computer.
7. After Windows 95 has restarted, click Start,
Windows Explorer and make a new folder on
your C: drive named Password. Choose the
Password directory and make a text file within it
called Password Test.
8. Right-click on the Password directory to display
the context-sensitive menu.
FIGURE 10.15
This Properties application for a shared directory uses
share-level security.
16.919-x CH10.i
384
8/28/98 1:48 PM
C h a p te r 1 0
Page 384
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
A P P LY Y O U R L E A R N I N G
sharing hand symbol replaces the folder symbol
for the shared directory.
and that you will have to re-share all of your folders again. If this is the case, click OK.
11. If you have another computer on the network,
browse the first computer in Network
Neighborhood to display the share name. The
share name Password is displayed under the
appropriate computer name.
4. Type the name of the server with the user
accounts or the domain name into the Obtain list
of users and groups from field. Windows 95
attempts to access the Windows NT or NetWare
server to obtain the users list.
12. Double-click on the share name Password. You
are prompted for the password.
5. Select the Configuration tab and choose File,
Print Sharing to display the File and Print
Sharing dialog box.
13. Enter read at the password prompt and choose
OK to display the directory contents.
14. Copy the Password Test file from the share to
your local hard drive. The file read will be successful.
15. Modify the file and try to copy it back. Then try
to delete the original in the shared directory.
Neither the file write nor the file delete will be
allowed.
6. Select both the I want to be able to give others
access to my files check box and the I want to be
able to allow others to print to my printer(s)
check box to enable others to access your printers
and files. Then choose OK to automatically
install File and Printer Sharing for Microsoft
Networks. You may be prompted for the location
10.6 Sharing a Directory Using User-Level
Security
Objective: Allow Windows 95 to access another
machine’s user accounts list and share a directory using
user-level security.
Estimated time: 15 minutes
1. From the Start menu, choose Settings, Control
Panels to display the Control Panels window.
2. Double-click on the Network icon to display the
Network dialog box.
3. Choose the Access Control tab and select Userlevel access control, as shown in Figure 10.16. If
sharing is already installed, you will be notified in
a dialog box that all of your shares will be lost,
FIGURE 10.16
The User-level access control option enables advanced
Windows 95 networking security.
16.919-x CH10.i
8/28/98 1:48 PM
Page 385
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
385
of the source files; if this is the case, type in the
location and click on OK.
7. Choose OK and restart the computer. Until you
have restarted, the new settings will not take
effect, and you will not be able to complete the
exercise.
8. After Windows 95 restarts, click Start, Windows
Explorer. Create a new directory called UserLevel
and then create a new text file in the folder
named UserTest. Now choose the directory.
Observe that the Password folder is no longer
shared, as the change to user-level security results
in the loss of all existing shares. (This doesn’t
wipe out the files or folders, but it does eliminate
the logical network path to them.)
9. Right-click on the UserLevel directory to display
its context-sensitive menu.
10. Select Sharing from the context-sensitive menu to
display the Sharing tab of the Properties application.
11. Type UserLevl for the share name and give the
Local Training group full-access privileges by
selecting the group and choosing Full Access.
Choose OK. Then choose OK again. The folder
symbol for the shared directory is replaced with
an icon of a folder held by a hand.
12. Log on to another computer on the network
using the username to which you gave full-access
permissions. Locate the share name UserLevl in
the Explorer by browsing the entire network. The
share name UserLevl is displayed under the
appropriate computer name.
13. Double-click on the share name UserLevl to display the directory contents.
NOTE
A P P LY Y O U R L E A R N I N G
Share Names Sharing the folder
UserLevel with the 8.3 compatible
name UserLevl enables DOS workstations on the network to properly
understand the name of the share.
Remember that if all machines on
your network are not capable of long
filename support, you should continue
to use network file-naming conventions that correspond to your network’s lowest common denominator.
14. Try to copy a file to the shared directory. The file
write is allowed.
10.7 Creating a Local Printer with
Windows NT
Objective: Create a locally installed printer on NT.
Estimated time: 20 minutes
1. Click Start, Settings. Then choose Printers to
open the Printers window.
2. Double-click on Add Printer to display the Add
Printer Wizard. As with many other administrative tasks, the process of creating and sharing a
printer has been streamlined and simplified by
the use of a “wizard,” a small program that leads
you through a particular task. Choose My
Computer and click Next.
3. The wizard asks you to specify the port or ports
to which the new printer should print. Choose
lpt1: and click Next.
16.919-x CH10.i
386
8/28/98 1:48 PM
C h a p te r 1 0
Page 386
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
A P P LY Y O U R L E A R N I N G
PO R TS
You can define multiple ports because printers in
Windows NT and Windows 95 are virtualized. This
refers to the fact that a printer in these environments
is not a physical machine, but rather a collection of
settings and configuration information about a particular machine. You can test this by installing a printer—
or a modem or a network card—that is not actually
physically present in your machine. The device will
install perfectly, but you will receive an error if an
application attempts to access it, because this device
doesn’t exist. When a matching physical device is
attached, the virtual device will recognize it and forward information as needed. For those of you who
have used print queues, it might be easier to think of
a printer as a new name for a print queue. The
machine itself is referred to as a printing device.
Finish. You will need the source files for both
Windows NT Server or Workstation and for
Windows 95. You are prompted for the location
of the source files, and the necessary drivers are
loaded.
8. The Printer icon for Color Printer is created in
the Printers window. Select it, and the queue
appears. Print a document to the new printer and
check the queue again. The document should be
waiting to print.
10.8 Sharing a Printer on the Network with
Windows 95
Objective: To share a printer on the network from
Windows 95.
Estimated Time: 5 minutes
4. The wizard now asks you to specify the type of
physical device to which you are printing or the
device type that your printer emulates. Click HP
in the left pane and then find and select HP
Color LaserJet. Then click Next.
5. Now you are asked to name your new printer.
Remember that each printer on your machine
must have a unique name, and that name should
be descriptive of its type or function. Type Color
Printer and click Next.
6. Now you are asked whether the printer will be
shared, and if so, what other operating systems
will access it. Click Shared, call the new share
MyLaser, and select Windows 95 from the list of
additional operating systems. Note that each supported Windows NT platform requires a different
driver. Click Next.
7. The wizard now has all the information it needs.
Leave the Print Test Page option on and click
This lab assumes that a printer is already installed.
1. Click My Computer and double-click on the
Printers folder. Right-click on the Printer icon
and choose Properties to display the Properties
sheet.
2. Choose the Sharing tab to display the Sharing
configuration settings.
3. Select Shared As and enter a share name and an
optional descriptive comment for the printer.
Windows 95 does not allow a share name to
contain invalid characters, including spaces. In
addition, the share name must not exceed 12
characters.
4. You also must grant permissions to access this
printer. If share-level permissions are used, you
must assign a password to the printer. To access
the print queue, users must supply the correct
password. If user-level permissions are used, you
16.919-x CH10.i
8/28/98 1:48 PM
Page 387
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
387
A P P LY Y O U R L E A R N I N G
must add the users who will be granted access to
this print queue. For example, to enable everyone
to print to the print queue, you would add the
Everyone group and grant it the print access
right. If you can, assign these permissions from
what you have learned. If you have problems,
refer back to previous exercises for instructions
on setting permissions. Remember that files and
printers are shared through the same process with
just a few twists.
5. Choose OK to share the printer. The Printer icon
now appears as a hand holding or sharing the
printer with others. Remote users with the correct permissions now can access the print queue
after setting up the correct printer driver on their
computers.
C. Local group
D. Container user
2. This group usually is used to assign permissions.
A. Global
B. LAN
C. Local
D. Either global or local
3. Share-level permissions enable which of the following actions?
A. Defining access levels by user
B. Controlling file-level access
C. Providing no access security at all
D. Defining access levels by password
Review Questions
1. What are the two types of security Windows 95
offers?
2. What is the difference between a primary domain
controller (PDC) and a backup domain controller (BDC)?
3. What is the difference between share-level security and file system security in Windows NT?
4. True or false: User-level access is less secure than
share-level access.
A. True
B. False
5. Which of the following does not add additional
security to your network?
A. Auditing
B. Virus scanning
Exam Questions
1. Which type of account is only available on
Windows NT domain controllers?
A. User
B. Global group
C. Data compression
D. Data encryption
16.919-x CH10.i
388
8/28/98 1:48 PM
C h a p te r 1 0
Page 388
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
A P P LY Y O U R L E A R N I N G
6. “Log on Locally” is one example of which of the
following?
A. Permissions
B. Rights
C. Once on every Windows NT-based machine
D. You should avoid creating users and use
global groups instead.
11. This is required for local file-level security.
C. Privileges
A. NTFS
D. Attributes
B. Share-level security
7. Read and Change are two types of which of the
following?
A. Permissions
B. Rights
C. User-level security
D. FAT
12. What two types of general groups can Windows
NT domains include?
C. Privileges
A. Local
D. Attributes
B. Domain
8. A shared printer is available to whom?
A. Everyone on the network
B. Only the person who shared it
C. Anyone with rights to the share
D. Anyone with permissions to the share
C. Global
D. Everyone
13. True or false: A single user can be placed in more
than one group.
A. True
B. False
9. Which group includes everyone who uses a
resource locally?
A. Everyone
B. Interactive
C. Creator-Owner
D. Network
10. How many times should each user be created in a
single domain?
A. Once
B. Once on each domain controller
14. You want to centralize the management of your
network.
Required result: You want to have passwords and
user accounts as part of your security.
Optional result 1: You want to have file system
security on your hard drive.
Optional result 2: You want to implement sharelevel security on your shared directories and
printers.
16.919-x CH10.i
8/28/98 1:48 PM
Page 389
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
389
A P P LY Y O U R L E A R N I N G
Suggested solution: You install Windows 95 and
implement share-level security. You make sure
that the hard drives have NTFS partitions.
A. This solution will obtain the required result
and both optional results.
B. This solution will obtain the required result
and one optional result.
C. This solution will obtain the required result.
D. This solution does not satisfy the required
result.
15. You want to centralize the management of your
network.
Required result: You want to have passwords and
user accounts as part of your security.
Optional result 1: You want to have file system
security on your hard drive.
Optional result 2: You want to implement share
level security on your shared directories and
printers.
Suggested solution: You install Windows NT and
implement share-level security. You make sure
that the hard drives have FAT partitions.
A. This solution will obtain the required result
and both optional results.
B. This solution will obtain the required result
and one optional result.
C. This solution will obtain the required result.
D. This solution does not satisfy the required
result.
16. You want to centralize the management of your
network.
Required result: You want to have passwords and
user accounts as part of your security.
Optional result 1: You want to have file system
security on your hard drive.
Optional result 2: You want to implement share
level security on your shared directories and
printers.
Suggested solution: You install Windows NT and
implement share-level security. You make sure
that the hard drives have NTFS partitions.
A. This solution will obtain the required result
and both optional results.
B. This solution will obtain the required result
and one optional result.
C. This solution will obtain the required result.
D. This solution does not satisfy the required
result.
Answers to Review Questions
1. Windows 95 offers two types of security: sharelevel and user-level.
Share-level security is available at all times to
Windows 95 to allow peer-to-peer networking.
Share-level security relies on the use of passwords
to protect individual shared resources. Each
resource would have its own password.
User-level security relies on the presence of a
Windows NT Workstation, Windows NT Server,
16.919-x CH10.i
390
8/28/98 1:48 PM
C h a p te r 1 0
Page 390
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
A P P LY Y O U R L E A R N I N G
or a Novell server to supply a database of user
accounts. User-level security provides security on
a user-by-user or group-by-group basis.
See the section titled “Implementing Security on
Windows 95.”
2. A primary domain controller (PDC) and a backup domain controller (BDC) provide essentially
the same functions. They both maintain a User
Account database as well as process logon
requests for users on client machines. The main
difference between a PDC and a BDC is that a
PDC contains the master copy of the user
account database. The BDCs contain a replica
copy of this master database. All changes to the
database—that is, the addition and deletion of
user accounts—are done on the PDC’s user
account database.
See the section titled “Implementing Security on
Windows NT.”
3. Share-level security is the ability to share a directory out on the network and assign this share permissions. The permissions available for this are
Full Control, Read, Change, and No Access.
These permissions only apply to users accessing
the share from a remote machine. Share-level
security can be applied to FAT and NTFS partitions.
File system security is a level of security that can
be applied on an NTFS partition. This level of
security applies to directories and files. File system security applies to people accessing these files
over the network, or if the user is interactive to
the PC.
See the section titled “Implementing Security on
Windows NT” and “Implementing Security on
Windows 95.”
Answers to Exam Questions
1. B. User accounts and local groups exist on all
Windows NT computers. A container user is a
fictitious term. See the section titled “Managing
User Accounts and Groups Using Windows NT.”
2. C. Local groups are normally used to assign permissions. The general practice is to place users
into global groups, and then place the global
groups into local groups. The local groups are
then assigned permissions. See the section titled
“Managing User Accounts and Groups Using
Windows NT.”
3. D. Share-level permissions are applied to files and
printers on Windows 95 computers. This level of
security does not allow for a user-by-user level of
security. Because of this, A and B are both incorrect. C is incorrect, as share-level security does
provide access security. See the section titled
“Share-level Security on Windows 95.”
4. B. This answer is false because share-level security
does not control access based upon who is accessing the share, but instead controls on whether a
person knows the password. See the section titled
“Implementing Security on Windows 95.”
5. C. Data compression allows you to obtain more
disk space on a hard drive. See the section titled
“Additional Administrative Tasks” in Chapter 9,
“Disaster Recovery.”
6. B. The ability to log on locally is a right.
Permissions deals with resource shares. Privileges
is just another word in the English language.
Attributes are what files and directories will allow
to be performed on them. See the section titled
“Managing User Accounts and Groups Using
Windows NT.”
16.919-x CH10.i
8/28/98 1:48 PM
Page 391
Ch a p te r 10
MANAGING AND SECU RING A MICROSOFT NETWO RK
391
A P P LY Y O U R L E A R N I N G
7. A. Read and Change are two types of
Permissions. Rights are what the system will
allow you to do to it, such as “log on locally” or
“back up files and directories.” Privileges is
another word in the English language that sounds
good. Attributes are what files and directories will
allow to be performed on them. See the section
titled “Managing User Accounts and Groups
Using Windows NT.”
8. D. Resource shares are based upon permissions.
See the answers for question 7, 8, and 9 for why
C is incorrect. A and B are incorrect, as these
would be dependent upon what permissions were
assigned to the share. See the section titled
“Permissions.”
9. B. The Interactive group is that special group
that applies to anyone locally on a machine.
Network group is anyone on a remote machine
from the resource. Creator-Owner is a file permission. Everyone is all people in the domain.
See the section titled “Groups.”
10. A. A user can only be created once in a single
domain. See the section titled “User Accounts.”
11. A. File-level security is only applied on Windows
NT computers that have an NTFS partition. D is
incorrect, as FAT does not support file-level security. B and C are incorrect, as they are Windows
95 security models. See the section titled
“Assigning File-Level Permissions on an NTFS
Partition.”
12. A, C. Windows NT domains have two general
types of groups, local and global. C is incorrect
because it is a specific group. B is incorrect
because there does not exist a domain group.
See the section titled “Groups.”
13. A. A user can be placed in as many groups as the
administrator wishes. See the section titled
“Groups.”
14. D. Windows 95 does not have the ability to create a user account database, nor can it provide file
system security with an NTFS partition. It can,
however, offer share-level security. The only possible result that could be correct would be optional
result 2. See the sections titled “Implementing
Security on Windows NT” and “Implementing
Security on Windows 95.”
15. B. Both the required result and the second
optional result are met with the proposed solution. The first optional result cannot be met, as
only NTFS partitions can offer file system security. See the section titled “Implementing Security
on Windows NT.”
16. A. All results are met with the installation of
Windows NT using NTFS partitions. See the
section titled “Implementing Security on
Windows NT.”
16.919-x CH10.i
392
8/28/98 1:48 PM
C h a p te r 1 0
Page 392
M ANAG I NG AND SECU RING A MICROSOFT NETWORK
Suggested Readings and Resources
1. Siyan, Karanjit S. Windows NT Server 4
Professional Reference. New Riders, 1996.
2. Heywood, Drew. Inside Windows NT Server 4.
New Riders, 1997.
3. Casad, Joe. MCSE Training Guide: Windows
NT Server 4. New Riders, 1997.
4. Sirockman, Jason. MCSE Training Guide:
Windows NT Server 4 Enterprise. New Riders,
1997.
5. Boyce, Jim. Inside Windows 95, Deluxe Edition.
New Riders, 1996.
17.919-x CH11.i
8/28/98 1:56 PM
Page 393
OBJECTIVES
Chapter 11 targets the following objective in the
Implementation section of the Networking Essentials
exam:
Select the appropriate hardware and software
tools to monitor trends on a given network.
. One of the most important things an administrator
can do is monitor the network. By monitoring the
network, the administrator is able to determine the
demand placed upon the system and the usage of
resources. This exam objective is designed to
encourage you to develop your ability to determine
what tools you should use in monitoring the network for trends.
C H A P T E R
11
Monitoring the
Network
17.919-x CH11.i
8/28/98 1:56 PM
Page 394
OUTLINE
S T U DY S T R AT E G I E S
Monitoring Network Trends
395
Keeping Records
396
Monitoring Performance
397
Simple Network Management Protocol
(SNMP)
397
Windows NT Performance Monitor
398
Windows 95 System Monitor
400
Monitoring Network Traffic
401
Logging Events
402
Chapter Summary
407
. When you read this chapter, be very aware of
the role that different devices and programs
play in monitoring the network. Pay particular
attention to the following Microsoft programs:
Performance Monitor, Network Monitor, System
Monitor, and Event Viewer. Be aware of what
each of these programs can monitor. Each of
these programs resides either on a Windows
95 or Windows NT computer, or both. Each of
these programs also has a particular set of
component of information that it records. By
becoming proficient in recognizing when you
should use which program, you should have no
difficulty in meeting the exam objective.
17.919-x CH11.i
8/28/98 1:56 PM
Page 395
C hapter 11
MONITORING THE NETWO RK
INTRODUCTION
An important part of network management involves monitoring
trends on the network. By effectively monitoring network behavior,
you can anticipate problems and correct them before they disrupt
the network. Monitoring the network also provides you with a baseline, a sampling of how the network functions in its equilibrium
state. By establishing a baseline on your system, you can determine
whether your network can handle the current resource usage or
whether additional resources are needed.
This chapter presents various programs or mechanisms that can be
used to monitor and record information about the network. The
explanation of what these different mechanisms are and when you
would utilize them is addressed in this chapter.
MONITORING NETWORK TRENDS
Monitoring the network is an ongoing task that requires data from
several different areas. Some of the monitoring tools that keep watch
on the network are discussed in other chapters. The purpose of this
chapter is to bring these tools together so that you can view them in
the context of an overall network monitoring strategy. The following
list details some tools you can use to document network activities:
á Written documentation
á A statistics-gathering or performance-monitoring tool, such as
Windows NT’s Performance Monitor
á A network-monitoring and protocol-analysis program—such
as Windows NT’s Network Monitor or the more powerful
Network Monitor tool included with Microsoft’s BackOffice
System Management Server (SMS) package—or a hardwarebased protocol analyzer
á A system event log, such as the Windows NT event log,
which you can access through Windows NT’s Event Viewer
application
<$Inetworks;monitoring tools;written
documentation><$I
networks;monitoring
tools;Performance
Monitor><$Inetwor
ks;monitoring
tools;Network
Monitor><$Inetwor
ks;monitoring
395
17.919-x CH11.i
396
8/28/98 1:56 PM
C h a p te r 1 1
Page 396
M O NI TO R I NG TH E NETWORK
KEEPING RECORDS
A detailed history of changes to the network serves as a tremendous
aid in troubleshooting. When a problem occurs, the first thing you
want to know is what has changed and when it was changed. This
information can be gathered from written documentation.
<$Inetworks;monitoring tools;written documentation><$Imonit
oring tools (networks);written documentation><$Irecordk
eeping;network monitoring><$Iadminstrat
ors;monitoring
Your documentation of the network should begin from the day the
network is installed. The layout, design, components, and software
should all be recorded within your network documentation. Contact
names, service contracts, as well as important support telephone
numbers should also be part of your network’s documentation.
This documentation can be as simple as sketched designs and information written on a piece of paper, or as elaborate as an electronic
schematic and database stored on a computer.
The following list details some items your configuration records
should include:
á Descriptions of all hardware, including installation dates,
repair histories, configuration details (such as interrupts and
addresses), and backup records for each server
á A map of the network showing locations of hardware compo-
nents and cabling details
á Documentation describing why certain layouts or naming con-
ventions were chosen, so that these conventions can be followed in the future
á Current copies of workstation configuration files, such as
CONFIG.SYS and AUTOEXEC.BAT files for DOS and
Windows 3.1 machines, or backup copies of the registry files
for Windows 95 and NT
á Service agreements and important telephone numbers, such as
the numbers of vendors, contractors, and software support
lines
á Software licenses to ensure that your network operates within
the bounds of the license terms
á A history of past problems and related solutions
17.919-x CH11.i
8/28/98 1:56 PM
Page 397
C hapter 11
Records of the network are used for more than just troubleshooting.
They also supply a wealth of information for future planning.
Records can help you maintain consistency within the hardware and
software. Detailed records also save a lot of time when software and
hardware audits are performed—a common event within mediumand large-sized organizations.
MONITORING THE NETWO RK
397
<$Inetworks;monitoring
tools;written documentation><$Imonitoring tools (networks);written documenta-
MONITORING PERFORMANCE
One of the most important tasks that should be performed on the
network is some form of statistical collecting. These statistics can
range from the performance of servers, workstations, and other
devices on the network to the performance of individual components within a program or service itself. This section looks at three
types of performance monitoring tools: Simple Network Management Protocol (SNMP), Windows NT Performance Monitor, and
Windows 95’s System Monitor.
Simple Network Management Protocol
(SNMP)
Many types of software and hardware on the market enable you to
collect statistics on the network. One important protocol used within the TCP/IP protocol suite that assists in statistic collecting is the
Simple Network Management Protocol (SNMP).
SNMP is a protocol that is supported by most pieces of hardware
and software that support the TCP/IP protocol stack. This protocol
allows for the collection of statistics of various resources on the network. For this information to be collected about a resource, the
resource must run an SNMP service, or have some other device run
the SNMP service on its behalf.
The SNMP service collects predefined information. This information is stored in a Management Information Base (MIB). An MIB is a
database of information that can be read by management software
designed to work with SNMP. An example of this management software is IBM’s OpenView.
<$Inetworks;monitoring tools;statistical collecting><$Imonitoring
tools (networks);statistical collec
ting><$Inetworks;monitoring
tools;SNMP><$Imonitoring tools
(networks);SNMP><$ISNMP;(Sim
ple Network Monitoring
Protocol)><$ISNMP;use within
TCP/IP protocol
suite><$ISNMP;management information base
(MIB)><$ISNMP;IBM OpenView
software><$IIBM OpenView software>
17.919-x CH11.i
398
8/28/98 1:56 PM
C h a p te r 1 1
Page 398
M O NI TO R I NG TH E NETWORK
<$Inetworks;monitoring
tools;SNMP><$Imonitoring tools
(networks);SNMP><$Imanagement
information base
(MIB);SNMP><$ISNMP;traps;issu
ing><$Itraps;issuing;SNMP>
Management software issues one of the following three main commands:
á The get command gathers information within an MIB.
á The get next command gets the next piece of information
within the MIB.
á The set command places information within the MIB.
These devices that have an SNMP service monitoring them can also
be configured to issue traps, or system messages, when certain parameters are reached or exceeded.
Windows NT Performance Monitor
<$Inetworks;monitoring
tools;Performance
Monitor><$Imonitor
ing tools (networks);Performance
Monitor><$IPerform
ance Monitor;counters><$IPerformance
Windows NT’s Performance Monitor tool lets you monitor important system parameters for the computers on your network in real
time. Performance Monitor can keep an eye on a large number of
system parameters, providing a graphical or tabular profile of system
and network trends. Performance Monitor also can save performance
data in a log for later reference. You can use Performance Monitor to
track statistical measurements (called counters) for any of several
hardware or software components (called objects). An example of
these counters for an object being displayed in a chart format can be
seen in Figure 11.1.
Some Performance Monitor objects that relate to network behavior
are as follows:
á Network segment
á Server
á Server work queues
á Workstation or other Redirectors
á Protocol-related objects, such as TCP, UDP IP, NetBEUI,
NWLink, and NetBIOS
á Service-related objects, such as Browser and Gateway Services
for NetWare
17.919-x CH11.i
8/28/98 1:56 PM
Page 399
C hapter 11
MONITORING THE NETWO RK
FIGURE 11.1
A Windows NT Performance Monitor chart.
Some Performance Monitor counters that relate to the performance
of components or resources on a computer are as follows:
á Processor
á Memory
á PhysicalDisk
Software installed on Windows NT can, in some cases, also create its
own Performance Monitor counters. Microsoft’s BackOffice products create a vast number of objects and counters for you to monitor. Third-party products can also create objects and counters to
monitor.
You should use Performance Monitor if you are experiencing problems, but you should also use Performance Monitor to log network
activity when things are running smoothly. Logging normal network
activity, especially after the network has been installed or a new
resource has been added or changed helps you establish a baseline.
It is this baseline to which later measurements can be compared.
Exercises 11.2 and 11.3 at the end of this chapter provide you with
a guided tour of Windows NT’s Performance Monitor application.
More information about Performance Monitor is covered within
<$IPerformance
Monitor;counters;chart diagram><$IPerfor
mance
Monitor;when
to
run><$IPerform
399
17.919-x CH11.i
400
8/28/98 1:56 PM
C h a p te r 1 1
Page 400
M O NI TO R I NG TH E NETWORK
<$IPerformance Monitor;hard
drive;command prompt>
courses dealing with Windows NT Server, Workstation, and the
Windows NT Enterprise books and courses.
One final thing to note is that Performance Monitor does not gather
information relating to the hard drive of a computer unless the command diskperf –y is run from the command prompt. Use diskperf –ye
if the hard drive is part of a stripe set. After these commands are run,
monitoring of the hard drives will continue until diskperf – n is
issued at the command prompt. If you issue any of these commands,
the computer must be rebooted for them to take effect.
<$Inetworks;monitoring
tools;System
Monitor><$Imoni
toring tools (networks);System
Monitor><$IWind
ows 95;System
Monitor><$ISyste
m Monitor;versus
Performance
Monitor><$ISyste
Windows 95 System Monitor
Windows 95 includes a program called System Monitor that also
enables you to collect information on the Windows 95 machine in
real time. System Manager collects information on different
Categories of Items on the system (see Figure 11.2).
The main categories within System Monitor are as follows:
á File System. Information written to or read from the hard drive
á IPX/SPX compatible protocol. Information on the number of
IPX and SPX packets sent out from and received by the computer
á Kernel. Processor usage, number of threads being processed,
and the number of virtual machines running on the computer
FIGURE 11.2
Windows 95’s System Monitor.
á Memory Manager. Various memory items that can be tracked
on the computer
á Microsoft Network Client. The number of files, sessions,
resources, and bytes sent or received by the network client
Although Windows 95’s System Monitor is not as elaborate or extensive in the collection of information and statistics as Windows NT’s
Performance Monitor, nor can System Monitor collect information
to a log file, it is a useful tool for baselining a Windows 95 computer.
In most cases a constant recording of system information by
Performance Monitor or System Monitor is not needed. The continuous use of these programs consumes computer resources that are
best left for other programs. Also, after a baseline is established, the
collection of redundant information is often not warranted.
17.919-x CH11.i
8/28/98 1:56 PM
Page 401
C hapter 11
MONITORING THE NETWO RK
401
MONITORING NETWORK TRAFFIC
Protocol analysis tools monitor network traffic by intercepting and
decoding frames. Software-based tools, such as Windows NT
Server’s Network Monitor (see Figure 11.3), analyze frames coming
and going, in real time, from the computer on which they run.
Network Monitor records a number of statistics, including the percent of network utilization and the broadcasts per second. In addition, Network Monitor tabulates frame statistics (such as frames sent
and received) for each network address.
An enhanced version of Network Monitor, which is included with
the Microsoft BackOffice System Management Server (SMS) package, monitors traffic on more than just the traffic between the local
computer and other devices. It will also monitor traffic that is just
between other devices, and also traffic on remote networks, provided
a monitor agent is installed on the remote network segment.
For large networks, or for networks with complex traffic patterns,
you might want to use a hardware-based protocol-analysis tool. A
hardware-based protocol analyzer is a portable device that can be as
small as a palmtop PC or as large as a suitcase. The advantage of a
hardware-based protocol analyzer is that you can carry it to strategic
places around the network (such as a network node or a busy
cabling intersection) and monitor the traffic at that point.
<$Inetworks;traffic;monitoring
tools><$Imonitoring tools
(networks);Network
Monitor><$Itraffic;monitoring (Network
Monitor)><$INetwork
Monitor;statistics><$Inetw
orks;hardware-based protocol analysis
tools><$INetwork
Monitor;traffic monitiring
capabilities>
FIGURE 11.3
Windows NT Server’s Network Monitor main
screen.
17.919-x CH11.i
402
8/28/98 1:56 PM
C h a p te r 1 1
Page 402
M O NI TO R I NG TH E NETWORK
<$Inetworks;traffic;monitoring tools><$Imonitoring
tools (networks);Network
Monitor><$Itraffic;monitoring (Network
Monitor)><$INetwork
Monitor;statistics><$Inetw
Some protocol analyzers are quite sophisticated. In addition to keeping network traffic statistics, they can capture bad frames and often
isolate the source. They also can help determine the cause of bottlenecks, protocol problems, and connection errors. A hardware-based
protocol analyzer is often a good investment for a large network
because it concentrates a considerable amount of monitoring and
troubleshooting power into a single, portable unit. For a smaller network, however, a hardware-based analyzer might not be worth the
initial five-figure expense because less expensive software-based products perform many of the same functions.
LOGGING EVENTS
<$Inetworks;monitoring tools;Event
Viewer><$Imonitori
ng tools (networks);Event
Viewer><$IEvent
Viewer;event
types;system><$IEve
nt Viewer;event
types;security><$IEv
ent Viewer;event
types;application><$Iapplication
Some operating systems, such as Windows NT, have the capability
to keep a running log of system events. That log serves as a record of
previous errors, warnings, and other messages from the system.
Studying the event log can help you find recurring errors and discover when a problem first appeared. The event log should also be
scanned on a regular basis to look for any indications of potential
problems.
Windows NT’s Event Viewer application provides you with access to
the event log. You can use Event Viewer to monitor the following
types of events:
á System events. Warnings, error messages, and other notices
describing significant system events. Examples of system log
entries include browser elections, service failures, and network
connection failures.
á Security events. Events tracked through Windows NT’s auditing
features. Refer to Chapter 10, “Managing and Securing a
Microsoft Network.”
á Application events. Messages from Win32 applications. If you’re
having a problem with an application, you can check the
application log for an application-related error or warning
message, provided the application is programmed to write to
the event log. Some NT services such as the JET database
engine used by WINS record their information in the application events log rather than the system log.
17.919-x CH11.i
8/28/98 1:56 PM
Page 403
C hapter 11
Event Viewer is part of the Windows NT Server Administrative
Tools group. To start Event Viewer, click on the Start button and
choose Programs, Administrative Tools, Event Viewer. Figure 11.4
shows the Event Viewer main screen. Click on the Log menu to
select the System, Security, or Application log.
If you double-click on a log entry in Event Viewer, an Event Detail
dialog box appears on your screen (see Figure 11.5). An Event Detail
provides a detailed description of the event.
The Windows NT Event Viewer utility contains the following five
types of events:
á Information. These events simply state that something of
importance has been done, such as the loading of a protocol.
These events are recorded for a matter of information only.
MONITORING THE NETWO RK
<$Inetworks;monitoring
tools;Event
Viewer><$Imonitoring tools
(networks);Event
Viewer><$IEvent
Viewer;launching><$IEvent
Viewer;event
types;Information><$IEvent
Viewer;event
types;Warning><$IEvent
Viewer;event
types;Stop><$IEvent
Viewer;event
á Warning. These events serve as a warning that some event that
may be important has occurred. Often when services are
stopped, a warning event is generated.
á Stop. These events occur when something of significance, such
as a detrimental event, has occurred. Often when services or
hardware fail, a Stop event is generated.
á Success. This event is generated within the auditing log.
Success events are generated when an object that was audited
as successful has occurred. You might, for example, audit the
successful logon of users.
á Failure. This event is generated within the auditing log. Failure
events are generated when an object that was audited as a
Failure has occurred, such as the failure of users to log on.
FIGURE 11.5
Event detail describing a system event.
FIGURE 11.4
The Event Viewer main screen.
403
17.919-x CH11.i
404
8/28/98 1:56 PM
C h a p te r 1 1
Page 404
M O NI TO R I NG TH E NETWORK
C A S E S T U DY : M O N I T O R I N G
THE
ESSENCE OF THE CASE
The essence of the case is as follows:
NETWORK
ing order: records and documentation, baselining, and monitoring.
• What documentation will be required?
• What baselining will be done?
• What preventive monitoring will be performed?
SCENARIO
Books Unlimited, a major up-and-coming retail
book chain, has decided to implement a network.
You have been called in to set up the network, as
well as tune the network, so that problems can
be avoided and a preventive maintenance program can be implemented.
Books Unlimited requires a general list of what
steps and measures you wish to implement on
the network for monitoring purposes. This information will be collected before and after the network is installed. This list will act as a guideline
for the eventual monitoring and maintenance program that will be implemented.
The Books Unlimited network will be a Windows
NT network, with Exchange, SQL, and SMS
BackOffice products being installed. The transport protocol to be used is TCP/IP. When possible, Books Unlimited would like to use resources
that come with NT, and SMS to monitor
resources in the network.
A N A LY S I S
To address the needs of Books Unlimited, the
three primary issues are addressed in the follow-
Records and Documentation
One of the things that should be performed on
any network, right from the beginning, is the collection of documentation and the keeping of
records. The Books Unlimited network should be
no exception.
Records and documentation begin before the first
components of a network are put into place. A
design of the Books Unlimited network physical
layout, file system structures, naming conventions, and hardware types should be put down on
paper. In most cases, hardware and software are
listed for cost purposes, but a proper design
shows how and why the design being implemented was used. A completely documented design
includes details such as why one cable type versus another was chosen, why one network card
versus another was selected, and even the reasons why the naming conventions being put into
place were chosen.
In addition to the paper information of the
design, components, and reasoning, all documentation that came with any piece of hardware
should also be kept. This includes items such as
warranties, licenses, and configuration information. Documentation for any software shipped
with the hardware should also be filed away for
future use.
All service contracts, licenses, warranties, purchase receipts, and contact numbers for the
Books Unlimited network should also be documented. This information is beneficial if the company ever needs to contact key people in the
17.919-x CH11.i
8/28/98 1:56 PM
Page 405
C hapter 11
C A S E S T U DY : M O N I T O R I N G
THE
event of an emergency, or if a corporate audit is
performed on the software and hardware of this
Books Unlimited location.
This collection of documentation and recordkeeping is something that will be continued for the
life of the network. As a network grows, new
components are being added all the time. The
addition of new components means new software
and documentation; therefore, keeping up-to-date
records is always important.
Baselining
Baselining of the Books Unlimited network is an
important first step, after hardware components
are put into place. In an ideal world, a person
would be able to develop a model of the future
network within a laboratory. But many companies
do not have the resources or time to perform
such research. Often, network designers need to
draw upon their experience and knowledge of networks to come up with the optimal solution.
In either case, after the network is actually
installed for use, it is very important to begin
baselining the network. Baselining is the recording of the resource usage on the network.
Baselining is done on a continual basis over several weeks, after a network is put into place.
Statistics such as processor usage or the speed
of the network can be recorded. Based upon the
information gathered, adjustments to the network
can be made. These adjustments could be as
simple as changing some parameters on some
of the computers to optimize resource usage, or
as involved as reevaluating the initial decisions
made on components such as cable types or
servers.
MONITORING THE NETWO RK
405
NETWORK
Microsoft’s Performance Monitor and Network
Monitor are good tools to monitor resources on
Books Unlimited’s network. Performance Monitor
basically tracks the components within a
Microsoft NT computer (System Monitor is used
for Windows 95), whereas Network Monitor
enables a person to see how well the usage of
the network is performing.
Microsoft Exchange, SQL, and SMS servers all
add objects and counters that can be monitored
with Performance Monitor. Careful research
should be done on each of these products to see
what boundaries or thresholds should be established by the resources being monitored. Also, if
you wish to monitor the TCP/IP protocol in
Performance Monitor, the SNMP service must be
loaded on the Windows NT computer.
After any new resource is added to Books
Unlimited’s network, a baseline of the usage or
performance of that resource should be undertaken. This ensures that the resource in question
is being used, or is configured, in its optimal
capacity.
Monitoring
In many ways, baselining is monitoring. The main
difference is that baselining is done after a new
network has been installed or a new resource
has been introduced to the network. Based upon
the information generated by the baseline and
saved by the network administrator, adjustments,
if needed, can be made to the new network or
added resource. This information on the network
that is saved is the baseline against which any
monitoring events in the future will be compared.
Monitoring, on the other hand, is the ongoing
continues
17.919-x CH11.i
406
8/28/98 1:56 PM
C h a p te r 1 1
Page 406
M O NI TO R I NG TH E NETWORK
C A S E S T U DY : M O N I T O R I N G
THE
NETWORK
continued
process of measuring the network’s performance.
Monitored events should be compared periodically to the baseline to see whether any abnormal
events are occurring.
from these utilities should be examined on a
daily basis. Often, a few minutes several times a
day is all that is needed to verify that all is well
on the network.
As time goes on, resources may be used in ways
not previously thought of by Books Unlimited.
Equipment begins to fail over time. Monitoring
the Books Unlimited network on a continual
basis enables administrators to detect resources
about to fail or the unexpected use of certain
resources.
A protocol analyzer is also a good tool to use for
analyzing the amount of network traffic being generated on the network. A protocol analyzer is also
a great tool to be used when there is some form
of connectivity issue. For example, if a person is
running an application on the server, and this
application is always failing at a certain point of
use, a protocol analyzer can detect and display
the actual file that is causing the application to
hang. In the case of Books Unlimited, they have
a small LAN only, and hence a software protocol
analyzer, as opposed to a hardware protocol analyzer, is probably a more cost-efficient alternative.
Performance Monitor can be used under these
situations, but is often not ideal for continual
monitoring because it does place a load on
resources of the computer that is running
Performance Monitor. For continual monitoring,
one should run Performance Monitor on a computer that is not going to be part of the monitored set of resources. Also, Performance
Monitor tends to generate too much detailed
information that is very similar to that collected
through the baseline. More commonly used tools
to perform ongoing monitoring are the Event
Viewer, and SNMP in a TCP/IP network. Both of
these utilities allow for the collection of events
on an ongoing basis. The information generated
In summary, recordkeeping, baselining, and monitoring of the network should always be done by
Books Unlimited. If all these tasks are done, the
network can be properly adjusted and perform at
its optimum. Also if a problem does arise, it can,
in most cases, be detected early by monitoring
and resolved quickly with the information contained in the network records and documentation.
17.919-x CH11.i
8/28/98 1:56 PM
Page 407
C hapter 11
MONITORING THE NETWO RK
CHAPTER SUMMARY
This chapter reviewed some of the tools and resources that can be
used to monitor network trends and information. These tools and
resources include the following:
á Documentation of the components placed on the network as
well as the network’s design
KEY TERMS
• Baseline
• Management Information Base
(MIB)
á Performance monitoring devices
• Performance Monitor
á Hardware- and software-based network monitoring and proto-
• System Monitor
col analysis tools
• Network Monitor
á Event logs
• Event Viewer
No single monitoring or recording method on its own is sufficient
on a network. You should instead use all the methods listed here.
• SNMP
<$IChapter 11;key terms><$Ikey
terms;Chapter 11>
407
17.919-x CH11.i
408
8/28/98 1:56 PM
C h a p te r 1 1
Page 408
M O NI TO R I NG TH E NETWORK
<$IChapter 11;exercises><$Iexercises;Chapter
11>
A P P LY Y O U R L E A R N I N G
Exercises
11.1 Using Network Monitor
Objective: Examine the main window display of
Windows NT Server 4.0’s Network Monitor application.
Estimated time: 15 minutes
1. If Network Monitor has been installed on your
system, click the Start menu and choose
Programs, Administrative Tools. Then choose the
Network Monitor application from the Administrative Tools group and proceed to Step 4.
2. If Network Monitor hasn’t been installed on your
system, you must install it, along with a component called the Network Monitor Agent. Network Monitor and the Network Monitor Agent
can be installed together with the Control Panel
Network application. Click the Start menu and
choose Settings, Control Panel. Double-click the
Network application and choose the Services tab.
3. In the Network application Services tab, click on
the Add button. Choose Network Monitor and
Agent from the Network Service list and click
OK. Windows NT prompts you for the
Windows NT installation disk. When the installation is complete, click OK and then Yes to shut
down your system and restart Windows NT.
Then start the Network Monitor application, as
described in Step 1.
4. Examine the four panes of the Network Monitor
main screen. The following list describes the four
panes:
• The Graph pane is located in the upper-left
corner of the display. The Graph section
includes five bar graphs describing network
activity. Only two of the graphs are visible;
use the scroll bar to view the other three
graphs.
• The Session Statistics pane, which appears
below the Graph pane, tracks network activity
by session, showing the two computers in the
session and the frames sent each way.
• The Total Statistics pane, which appears to
the right of the Graph pane, lists such important statistics as the number of frames and the
number of broadcasts. You can use the scroll
bar to reach other entries that are not visible.
• The Station Statistics pane, which sits at the
bottom of the window, shows statistics for
frames listed by network address.
5. Pull down the Capture menu and choose Start.
Network Monitor then starts monitoring the network.
6. Ping the Network Monitor PC from another
computer on the network. (Go to the command
prompt and type ping, followed by the IP
address on the Network Monitor computer—for
example, ping 111.121.131.141.) Watch the
Station Statistics pane at the bottom of the screen
to see any new information.
7. Experiment with sending files or other requests to
or from the Network Monitor PC. Study the
effect of network activity on the values displayed
in the four panes of the Network Monitor main
window.
8. When you are finished, pull down the Capture
menu and click Stop to stop capturing data.
Then exit Network Monitor. When prompted to
save your captured data, click on No.
17.919-x CH11.i
8/28/98 1:56 PM
Page 409
<$IChapter
11;exercises><$Iexercises;Chapter 11>
C hapter 11
MONITORING THE NETWO RK
409
A P P LY Y O U R L E A R N I N G
11.2 Creating a Chart in Performance
Monitor
Objectives: Become familiar with the process of creating and reading a Performance Monitor chart.
Understand the basic components of the Performance
Monitor main window and the Add to Chart dialog
box. Learn how to turn on disk performance counters
by using the diskperf command.
Estimated time: 25 minutes
1. From the Start menu, select Programs. Choose
the Administrative Tools group and click
Performance Monitor. The Performance Monitor
main window appears on your screen.
2. Pull down the Edit menu and choose Add to
Chart (see Figure 11.6). The Add to Chart dialog
box appears (see Figure 11.7). You can also
invoke the Add to Chart dialog box by clicking
the plus sign in the tool bar of the Performance
Monitor main window.
3. The box labeled Computer at the top of the Add
to Chart dialog box tells Performance Monitor
FIGURE 11.6
The Performance Monitor main window.
FIGURE 11.7
The Add to Chart dialog box.
which computer you want to monitor. The
default is the local system. Click the ellipsis button to the right of the box for a list of computers
on the network.
4. The box labeled Object tells Performance
Monitor which object you want to monitor. As
you learned earlier in this chapter, an object is a
hardware or software component of your system.
You can think of an object as a category of system
statistics. Pull down the Object menu. Scroll
through the list of objects and look for the
Processor, Memory, PhysicalDisk, LogicalDisk,
Server, and Network Segment objects described
earlier in this chapter. Choose the PhysicalDisk
object. If you have more than one physical
disk on your system, a list of your physical disks
appears in the Instance box to the right of the
Object box. The Instance box lists all instances of
the object selected in the Object box. If necessary,
choose a physical disk instance.
5. The box labeled Counter displays the counters
(the statistical measurements) that are available
for the object displayed in the Object box. Scroll
through the list of counters for the PhysicalDisk
object. If you feel like experimenting, select a
17.919-x CH11.i
8/28/98 1:56 PM
Page 410
<$IChapter 11;exercises><$Iexercises;Chapter 11>
410
C h a p te r 1 1
M O NI TO R I NG TH E NETWORK
A P P LY Y O U R L E A R N I N G
different object in the Object box. Notice that
the new object is accompanied by a different set
of counters. Switch back to the PhysicalDisk
object and choose the %Disk Time counter.
Click the Explain button on the right side of the
Add to Chart dialog box. Notice that a description of the %Disk Time counter appears at the
bottom of the dialog box. To add this to your
chart, click on the Add button.
6. Click the Done button in the Add to Chart dialog box. The dialog box disappears, and you see
the Performance Monitor main window.
7. In the Performance Monitor main window, a vertical line sweeps across the chart from left to
right. You may also see a faint colored line at the
bottom of the chart recording a %Disk Time
value of 0. If so, you haven’t enabled the disk performance counters for your system. (If the disk
performance monitors are enabled on your system, you should see a spikey line that looks like
the readout from an electrocardiogram. You’re
done with this step. Go on to Step 8. If you still
have no disk activity, activate the help on the
start menu, and perform a maximum search. This
should generate a lot of disk activity.)
If you need to enable the disk performance counters, click the Start button and go to the command prompt. Enter the command diskperf -y.
Then reboot your system and repeat Steps 1–7.
(You don’t have to browse through the object and
counter menus this time.)
8. You should now see a spikey line representing the
percent of time that the physical disk is busy
reading or writing. Select Add to Chart from the
Edit menu. Select the PhysicalDisk object and
choose the counter Avg. Disk Queue Length.
Click the Add button. Then choose the counter
Avg. Disk Bytes/Read. Click the Add button and
then click the Done button.
9. Examine the Performance Monitor main window.
All three of the counters you selected should be
tracing out spikey lines on the chart. Each line
is a different color. At the bottom of the window
is a table showing which counter corresponds
with which color. The table also gives the scale of
the output, the instance, the object, and the computer.
10. Below the chart (but above the table of counters)
is a row of statistical parameters labeled: Last,
Average, Min, Max, and Graph Time. These
parameters pertain to the counter that is selected
in the table at the bottom of the window. Select a
different counter and you see that some of these
values change. The Last value is the counter value
over the last reading. Graph time is the time it
takes (in seconds) for the vertical line that draws
the chart to sweep across the window. If you have
several lines on the chart, Ctrl + H highlights different lines on the chart.
11. Start Windows Explorer. Select a file (a graphics
file or a word processing document) and choose
Copy from Explorer’s Edit menu. (This copies
the file you selected to the clipboard.) Go to
another directory and select Paste from the Edit
menu. (This creates a copy of the file in the second directory.) Minimize Explorer and return to
the Performance Monitor main screen. The disk
activity caused by your Explorer session is now
reflected in the spikes of the counter lines.
12. Pull down the Options menu and select Chart.
The Chart Options dialog box appears on your
screen (see Figure 11.8). The Chart Options dialog box provides a number of options governing
the chart display. The Update Time section
17.919-x CH11.i
8/28/98 1:56 PM
Page 411
<$IChapter
11;exercises><$Iexercises;Chapter 11>
C hapter 11
MONITORING THE NETWO RK
411
A P P LY Y O U R L E A R N I N G
enables you to choose an update interval. The
update interval tells Performance Monitor how
frequently it should update the chart with new
values. (If you choose the Manual Update option,
the chart updates only when you press Ctrl + U
or click Update Now in the Options menu.)
Experiment with the Chart Options or click the
Cancel button to return to the main window.
13. Pull down the File menu. Choose Exit to exit
Performance Monitor. Note that the Save Chart
Settings and Save Chart Settings As options in
the File menu enable you to save the collection of
objects and counters you’re using now so you can
monitor the same counters later and avoid setting
them up again. The Export Chart option enables
you to export the data to a file that you can then
open with a spreadsheet or database application.
The Save Workspace option saves the settings for
your chart, as well as any settings for alerts, logs,
or reports specified in this session. Learn more
about alerts, logs, and reports in Exercise 11.3.
11.3 Performance Monitor Alerts, Logs, and
Reports
Objectives: Become familiar with the alternative views
(Alert view, Log view, and Report view) available
through the Performance Monitor View menu. Log
performance data to a log file.
Estimated time: 25 minutes
1. Click Programs in the Start menu and choose
Performance Monitor from the Administrative
Tools group. The Performance Monitor main
window appears onscreen.
2. Pull down the View menu. You see four options,
as follows:
• The Chart option plots the counters you
select in a continuous chart (refer to
Exercise 11.1).
• The Alert option automatically alerts a network official or executes an application if the
predetermined counter threshold is surpassed.
• The Log option saves your system performance data to a log file.
• The Report option displays system performance data in a report format.
FIGURE 11.8
The Chart Options dialog box.
The setup is similar for each of these view formats. All use some form of the Add to Chart dialog box (refer to Exercise 11.2). All have options
that are configured through the first command at
the top of the Options menu. (The first command at the top of the Options menu changes its
name depending on the active view. It was the
Chart option in Exercise 11.2.)
3. Click the Alert option in the View menu.
17.919-x CH11.i
412
8/28/98 1:56 PM
C h a p te r 1 1
Page 412
M O NI TO R I NG TH E NETWORK
<$IChapter 11;exercises><$Iexercises;Chapter
11>
A P P LY Y O U R L E A R N I N G
4. Click the plus sign in the toolbar or choose Add
to Alert from the Edit menu. The Add to Alert
dialog box (see Figure 11.9) is similar to the Add
to Chart dialog box except for two additional
items at the bottom of the screen. The Alert If
box enables you to type in a threshold for the
counter. The Over/Under option buttons specify
whether you want to receive an alert if the
counter value is over or under the threshold
value. The Run Program on Alert box lets you
specify a command line that executes if the
counter value reaches the threshold you specify in
the Alert If box. You can ask Performance
Monitor to send a message to your beeper, to
send you an email message, or to notify your paging service.
Don’t specify a batch file in the Run Program on
Alert box. Performance Monitor uses Unicode
format, which can confuse the command-prompt
interpreter. (The < and > symbols, which are used
in Unicode format, are interpreted as a redirection of input or output.)
5. The default object in the Add to Alert dialog box
should be the Processor object. The default
counter should be %Processor Time. Enter the
value 5% in the Alert If box and make sure the
Alert If option button is set to Over. In the Run
Program on Alert box, type SOL. Set the Run
Program on Alert option button to First Time.
This configuration tells Performance Monitor to
execute Windows NT’s Solitaire program when
the %Processor Time exceeds 5%.
If the Run Program on Alert option button is not
set to First Time, Performance Monitor executes a
new instance of Solitaire every time the
%Processor Time exceeds 5%, which happens
every time it executes a new instance of Solitaire.
You’ll probably have to close Performance
Monitor using the X button or reboot to stop the
incessant shuffling and dealing.
6. Click the Add button and then click the Done
button. The Alert Legend at the bottom of the
Alert window describes the active alert parameters. The Alert Log shows every instance of an
alert (see Figure 11.10).
7. Make some changes to your desktop. (Hide or
reveal the task bar, change the size of the
Performance Monitor window—anything that
causes a 5% utilization of the processor.) The
FIGURE 11.9
FIGURE 11.10
The Add to Alert dialog box.
The Performance Monitor alert log.
17.919-x CH11.i
8/28/98 1:56 PM
Page 413
<$IChapter 11;exercises><$Iexercises;Chapter 11>
C hapter 11
MONITORING THE NETWO RK
413
A P P LY Y O U R L E A R N I N G
Solitaire program should miraculously appear on
your screen. In a real alert situation, Performance
Monitor executes an alert application rather than
starting a card game.
8. Pull down the Edit menu and select Delete Alert.
9. Pull down the View menu and select Log.
Performance Monitor’s Log view saves performance data to a log file rather than displaying it
on the screen.
10. Pull down the Edit menu and select Add to Log.
Notice that only the objects appear in the Add
to Log dialog box. The counters and instances
boxes don’t appear because Performance Monitor
automatically logs all counters and all instances of
the object to the log file. Select the Memory
Object and click Add. If you want, you can select
another object, such as the Paging File object,
and click Add again. When you are finished
adding objects, click Done.
11. Pull down the Options menu and select Log. The
Log Options dialog box appears on your screen
(see Figure 11.11). The Log Options dialog box
enables you to designate a log file that
Performance Monitor is to use to log the data. In
the File name box, enter the name exer2.log. You
also can specify an update interval. The update
interval is the interval at which Performance
Monitor records performance data to the log.
The Manual Update option button specifies that
the file won’t be updated unless you press Ctrl +
U or select Update Now from the Options menu.
Click the Start Log button to start saving data to
the log. Wait a few minutes and then return to
the Log Options dialog box and click the Stop
Log button.
12. Pull down the View menu and switch to Chart
view.
FIGURE 11.11
The Log Options dialog box.
13. Pull down the Options menu and select Data
From. The Data From dialog box enables you to
specify a source for the performance data that
appears in the Chart. Note that the default source
is Current Activity. (That is why the chart you
created in Exercise 11.2 took its data from current system activity.) The alternative to the
Current Activity option is to use data from a log
file. Click the Log File option button. Click the
ellipsis button to the right of the log file window
and select the exer2 file you created in Step 11.
Click OK.
14. Pull down the Edit menu and click Add to Chart.
Click the down arrow to the right of the Object
menu. Notice that your only object choices are
the Memory object and any other objects you
selected in Step 10. Select the Memory object.
Browse through the counter list and select
Pages/sec. Click the Add button. Select any other
memory counters you want to display and click
the Add button. Click Done.
15. The log file’s record of the counters you
selected in Step 14 appears in the chart in the
17.919-x CH11.i
414
8/28/98 1:56 PM
C h a p te r 1 1
Page 414
M O NI TO R I NG TH E NETWORK
<$IChapter 11;exercises><$Iexercises;Chapter
11><$IChapter 11;review questions><$Ireview
questions;Chapter 11><$IChapter 11;exam ques-
A P P LY Y O U R L E A R N I N G
Performance Monitor’s main window. Notice
that, unlike the chart you created in Exercise
11.1, this chart does not continuously sweep out
new data. That is because this chart represents
static data from a previous, finite monitoring
session.
16. Pull down the Edit menu and select Time
Window. The Time Window enables you to focus
on a particular time interval within the log file
(see Figure 11.12). In this example (because you
collected data for only a few minutes), the Time
Window may seem unnecessary. If you collected
data for a longer period, however, and you want
to zero in on a particular event, the Time
Window can be very useful. Set the beginning
and end points of your time window by adjusting
the gray start and stop sliders on the Time
Window slide bar. The Bookmark section at the
bottom of the dialog box enables you to specify a
log file bookmark as a start or stop point. (You
can create a bookmark by selecting the Bookmark
option from the Options menu while you are collecting data to the log file or by clicking the book
in the Performance Monitor tool bar.) Click OK
to view the data for the time interval.
17. Pull down the View menu and switch to Report
view. Pull down the Options menu and select
Data From. Switch the Data From setting back to
Current Activity. Report view displays the performance data in a text report rather than in a
graphics format.
18. Select Add to Report from the Edit menu. Select
the processor object and choose the %Processor
Time, %Interrupt Time, and Interrupts/sec counters. (Hold down the Ctrl key to select all three
and then click Add.) Select the PhysicalDisk
object and choose the %Disk Time, Avg. Disk
Queue Length, and Current Disk Queue Length
counters. Click the Add button. Select the
Memory object and choose the Pages/sec, Page
Faults/sec, and Available Bytes counters. Click the
Add button. Click Done.
19. Examine the main report window. Performance
Monitor displays a report of the performance
data you specified in a hierarchical format, with
counters listed under the appropriate object.
20. Select Exit in the File menu to exit Performance
Monitor.
Review Questions
1. What is the purpose of Simple Network
Management Protocol (SNMP)?
2. What is a protocol analyzer, and what are the two
main types?
3. What is the purpose of creating a baseline?
FIGURE 11.12
The Performance Monitor Input Log File Timeframe dialog
box.
4. What tools does Microsoft Windows NT and
Windows 95 offer to enable you to monitor the
network?
17.919-x CH11.i
8/28/98 1:56 PM
Page 415
<$IChapter 11;exam questions><$Iexam questions;Chapter 11>
C hapter 11
MONITORING THE NETWO RK
415
A P P LY Y O U R L E A R N I N G
Exam Questions
1. What is an advantage of hardware-based network
monitoring tools over software-based tools?
5. Which utility keeps a record of the repair histories of network hardware?
A. Network Monitor
A. They are less expensive.
B. Event log
B. They are easier to use.
C. Client Manager
C. A hardware-based tool can also serve as a PC.
D. Nothing—you must do it yourself
D. None of the above.
2. Which tool would you use to determine whether
a Windows NT Server system displayed the same
error message at the same time every day?
A. Network Monitor
6. What does an MIB stand for?
A. Management Information Base
B. Message Information Base
C. Managed Information BIOS
D. Motherboards Information Board
B. Performance Monitor
C. Event Viewer
D. None of the above
7. Which program comes only with Windows 95?
A. Network Monitor
B. Performance Monitor
3. Which tool would you use to determine whether
a Windows NT Server machine has enough
RAM?
A. Network Monitor
B. Performance Monitor
C. Event Viewer
D. None of the above
4. An enhanced version of Network Monitor is
included with which product?
A. Windows NT
B. Windows 95
C. SMS
D. SNMP
C. SNMP
D. System Monitor
8. You wish to use the correct utility to monitor the
network. All your computers are either Windows
NT or Windows 95.
Primary objective: You need to see the network
traffic that is generated between computers.
Secondary objective: You need to be able to log
performance information on your Windows NT
servers.
Secondary objective: You need to plot Windows
NT Performance information to charts.
Suggested Solution: You install Network Monitor
to trace the network traffic, and System Monitor
on the Windows NT computers to log and chart
performance information.
17.919-x CH11.i
416
8/28/98 1:56 PM
C h a p te r 1 1
Page 416
M O NI TO R I NG TH E NETWORK
<$IChapter 11;exam questions><$Iexam
questions;Chapter 11><$IChapter
11;answers><$Ianswers;Chapter 11>
A P P LY Y O U R L E A R N I N G
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objective.
D. This solution does not meet the primary
objective.
9. You wish to use the correct utility to monitor the
network. All your computers are either Windows
NT or Windows 95.
Primary objective: You need to see the network
traffic that is generated between computers.
Secondary objective: You need to be able to log
performance information on your Windows NT
servers.
Secondary objective: You need to plot Windows
NT Performance information to charts.
Suggested Solution: You install Network Monitor
on a Windows 95 machine to trace the network
traffic, and Performance Monitor on the Windows 95 computers to log and chart performance
information on the Windows NT computers.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
Secondary objective: You need to be able to log
performance information on your Windows NT
servers.
C. This solution meets the primary objective.
Secondary objective: You need to plot Windows
NT Performance information to charts.
D. This solution does not satisfy the primary
objective.
Suggested Solution: You install Performance
Monitor on all Windows NT machines to track
the network traffic and log and chart performance information.
A. This solution meets the primary objective and
both secondary objectives.
B. This solution meets the primary objective and
one secondary objective.
C. This solution meets the primary objective.
D. This solution does not meet the primary
objective.
10. You wish to use the correct utility to monitor the
network. All your computers are either Windows
NT or Windows 95.
Primary objective: You need to see the network
traffic that is generated between computers.
Answers to Review Questions
1. The purpose of SNMP is to record and deliver
statistics on various network parameters. This can
be information such as whether a buffer is
becoming too full, the IP address of a computer,
or the routing table within a router.
SNMP information is delivered to a management
utility such as SMS in a process known as a trap.
Information can also be obtained by a management utility, with the issuance of a get or get next
command. Information can be set with the
issuance of a set command.
See the section titled “Simple Network
Management Protocol (SNMP)” for more information.
17.919-x CH11.i
8/28/98 1:56 PM
Page 417
<$IChapter 11;answers><$Ianswers;Chapter 11>
C hapter 11
MONITORING THE NETWO RK
417
A P P LY Y O U R L E A R N I N G
2. The purpose of a protocol analyzer is to capture
and record network traffic being sent across the
transmission media.
Protocol analyzers are of two types: software
and hardware. Software programs are installed
into computers. They use the computer’s network connectivity to collect the data on the
transmission media. A hardware protocol analyzer
is a dedicated, portable device used for the collection of network information. It uses a built-in
adapter to connect to the transmission media.
See the section titled “Monitoring Network
Traffic” for more information.
3. The purpose of creating a baseline is to establish
a historic record of how a network and its
resources operate in equilibrium.
See the section titled “Introduction” for more
information.
4. Microsoft Windows NT comes with the following tools to collect and monitor information:
• Performance Monitor. This utility is used to
collect information about different objects
and counters on a Windows NT computer.
It collects information on a real-time basis.
• Event Viewer. This utility is used to collect
information on any events that have happened on a Windows NT system. It collects
information on events as they occur.
• Network Monitor. This utility is used to collect information that is going across the transmission media. It collects information in real
time.
See the sections titled “Windows NT Performance Monitor,” “Monitoring Network Traffic,”
and “Logging Events” for more information.
Microsoft Windows 95 comes with the following
utility:
• System Monitor. This collects information on
various resources installed on a Windows 95
computer. This utility collects information in
real time.
Answers to Exam Questions
1. D. Hardware-based network monitoring tools are
almost always more expensive and more complex
to use than software tools. These hardware tools
do not serve as PCs. They do, however, provide
more information for analysis and tend to be
portable. See the section titled “Monitoring
Network Traffic.”
2. C. Event Viewer logs error messages and the
times that they occurred. Network Monitor captures the message, if it is a network error, but not
the time. Performance Monitor does not track
messages. See the section titled “Logging Events.”
3. B. Performance Monitor monitors components of
an NT machine, such as RAM. Network Monitor
does not monitor RAM, and Event Viewer tracks
system messages, not hardware. See the section
titled “Monitoring Performance.”
4. C. The enhanced version comes with SMS. See
the section titled “Keeping Records.”
5. D. Repair histories are kept only if you document
them. See the section titled “Monitoring Network
Traffic.”
6. A. An MIB is a Management Information Base.
See the section titled “Simple Network
Management Protocol (SNMP).”
17.919-x CH11.i
418
8/28/98 1:56 PM
C h a p te r 1 1
Page 418
M O NI TO R I NG TH E NETWORK
<$IChapter 11;answers><$Ianswers;Chapter
11>
A P P LY Y O U R L E A R N I N G
7. D. Network Monitor and SNMP can be installed
on Windows 95 and NT; Performance Monitor
can be installed on only Windows NT. Only
System Monitor comes with Windows 95. See
the section titled “Monitoring Network Traffic”
and “Simple Network Management Protocol
(SNMP).”
8. C. The primary objective is met because Network
Monitor shows the network traffic generated
between two computers. Neither secondary
objective is met, because System Monitor can be
installed only on Windows 95 computers. See the
sections titled “Windows NT Performance
Monitor,” “Windows 95 System Monitor,” and
“Monitoring Network Traffic.”
charting of performance information on a
Windows NT computer. The primary objective is
not met because you would need Network
Monitor to see the network traffic generated by
two computers. See the sections titled “Windows
NT Performance Monitor,” “Windows 95 System
Monitor,” and “Monitoring Network Traffic.”
10. C. The primary objective is met because Network Monitor shows network traffic between
computers. Neither optional result can be met,
because Performance Monitor cannot be run on
Windows 95 computers. See the sections titled
“Windows NT Performance Monitor,” “Windows 95 System Monitor,” and “Monitoring
Network Traffic.”
9. D. Both secondary objectives are met because
Performance Monitor allows for the logging and
Suggested Readings and Resources
1. Siyan, Karanjit S. Windows NT Server 4
Professional Reference. New Riders, 1996.
2. Heywood, Drew. Inside Windows NT Server 4.
New Riders, 1997.
3. Casad, Joe. MCSE Training Guide: Windows
NT Server 4. New Riders, 1997.
4. Sirockman, Jason. MCSE Training Guide:
Windows NT Server 4 Enterprise. New Riders,
1997.
5. Boyce, Jim. Inside Windows 95, Deluxe Edition.
New Riders, 1996.
18.919-x Part4
8/28/98 2:05 PM
Page 419
P A R T
TROUBLESHOOTING
12 Troubleshooting
IV
18.919-x Part4
8/28/98 2:05 PM
Page 420
19.919-x CH12.i
8/28/98 2:07 PM
Page 421
OBJECTIVES
Chapter 12 targets the following objectives in the
Troubleshooting section of the Networking Essentials
exam:
Identify common errors associated with components required for communications
. The purpose of this test objective is to make sure
you are able to isolate what problems are associated
with what components on the network.
Diagnose and resolve common connectivity problems with cards, cables, and related hardware
. This exam objective reflects the need for you to be
able to not only diagnose common connectivity
problems with cards, cables, and other related hardware, but also to be able to resolve these problems
in order to reestablish connectivity on the network.
Resolve broadcast storms
. This exam objective is designed to ensure that you
understand what causes a broadcast storm and
methods of resolving these broadcast storms.
Identify and resolve network performance problems
. This exam objective addresses your ability to use
tools and understand issues relating to identifying
and resolving poor performance on a network.
C H A P T E R
12
Troubleshooting
19.919-x CH12.i
8/28/98 2:07 PM
Page 422
OUTLINE
S T U DY S T R AT E G I E S
Intiating the Troubleshooting Process
423
Using Troubleshooting Tools
424
Protocol Analyzers
425
Digital Volt Meter (DVM)
425
Time-Domain Reflectometers
425
Oscilloscope
425
Other Tools
426
Troubleshooting Transmission Media
and Other Network Components
426
Troubleshooting Cables and Connectors
427
Troubleshooting Network Adapter Cards
429
Troubleshooting HUBs and MSAUs
430
Troubleshooting Modems
431
Repeaters, Bridges, and Routers
432
Handling Broadcast Storms
433
Troubleshooting Protocols
434
NetBEUI
434
NWLink (IPX/SPX)
434
TCP/IP
435
Troubleshooting Network Performance
435
Handling Other Network Problems
436
Getting Support
437
Chapter Summary
440
. There is a lot of information in this chapter. One
method of studying this information would be to
memorize everything. This can be time consuming and may not help when you encounter novel
situations. Instead, take time to try to relate the
information on troubleshooting to all of the previous chapters.
. Pay attention to the devices that can be used to
troubleshoot network problems, particularly
when you would use a particular device.
19.919-x CH12.i
8/28/98 2:07 PM
Page 423
C hapter 12
INTRODUCTION
Troubleshooting is the art of seeking out the cause of a problem and
eliminating the problem by managing or eliminating the cause.
With something as complex as a computer network, the list of possible problems and causes is nearly endless. In real life, however, a
large number of network problems fall into a few well-defined categories. One thing to be aware of is that no matter what the problem
is on your network, the OSI model serves as an excellent reference
tool to help you isolate the area of trouble. In this chapter, you learn
these categories. You will also learn about some of the strategies and
tools you can use to troubleshoot network problems.
Of course, no matter how effective you are at problem solving, it
almost always is better to avoid problems than to solve them.
Chapter 10, “Managing and Securing a Microsoft Network,”
discusses administration strategies that minimize the need for
troubleshooting. Chapter 11, “Monitoring the Network,” discusses
monitoring and record-keeping strategies that can help you identify
problems when they appear. This chapter looks specifically at troubleshooting techniques for solving problems related to network
cabling, adapter cards, modems, and other important connectivity
components. In addition, you learn some guidelines for troubleshooting network performance problems and are provided with
some important sources for finding troubleshooting information.
INITIATING
PROCESS
THE
TROUBLESHOOTING
Microsoft recommends the following five-step approach to network
troubleshooting:
1. Set the problem’s priority. Ask yourself a few questions: How
serious is this problem? Will the network still function if I
attend to other matters first? Can I quantify the loss of work
time or productivity the problem is causing? These will help
you determine the severity of the problem relative to the other
pressing problems you might face.
TROU BLESHOOT IN G
423
19.919-x CH12.i
424
8/28/98 2:07 PM
C h a p te r 1 2
Page 424
TRO UBL E S H O OTI NG
2. Collect information to identify the symptoms. Collecting
information can be as simple as asking users to describe the
problem in detail. A user’s description of the problem can lead
to further questions, which can lead to a deeper description. If
you keep a documented history of your network (see Chapter
11), you can compare the present behavior of the network
with the baseline behavior. You also can look for possible previous occurrences of the problem, and their documented solutions.
3. Develop a list of possible causes. Again, ask yourself a few
questions: Could the problem be related to connectivity
devices? Cabling? Protocols? A faltering workstation? What do
past occurrences have in common with the present occurrence?
List all possibilities.
NOTE
4. Test to isolate the cause. Develop tests that will prove or disprove each of the possible causes. The tests could be as simple
as checking a setup parameter or as complicated as studying
network traffic with a protocol analyzer. You learn about some
of the hardware and software network testing tools in the section titled “Using Troubleshooting Tools” later in this chapter.
Users in the Troubleshooting
Equation Also remember that if the
user is the problem, education and
training for the user is the fix.
5. Study the results of the test to identify a solution. Your tests
will (ideally) point you to the real problem; after you know the
problem, you can determine a solution. Make sure that when
testing, you perform only one change at a time, thereby eliminating the problem of determining what change provided the
solution.
These five steps are sufficient to guide you through a myriad of network problems. Similar approaches appear in the documentation of
other network vendors.
Part of the challenge of network troubleshooting is to determine
how you can apply these five troubleshooting steps to your own situation.
USING TROUBLESHOOTING TOOLS
Network administrators use a number of tools for searching out network problems. The following list details some of these tools. The
19.919-x CH12.i
8/28/98 2:07 PM
Page 425
Protocol Analyzers
EXAM
tools most commonly employed are Protocol Analyzers, Digital
Voltmeters, Time Domain Reflectometers, and Oscilloscopes.
TIP
C hapter 12
TROU BLESHOOT IN G
425
Troubleshooting Tools Although
these tools do not fall nicely into
any exam objective category, you
will be tested in your understanding
of what each of these tools does.
Protocol analyzers are hardware or software products that are used to
monitor network traffic, track network performance, and analyze
packets. Protocol analyzers can identify bottlenecks, protocol problems, and malfunctioning network components (see Chapter 11).
Digital voltmeters are hand-held electronic measuring tools that
enable you to check the voltage of network cables. They also can be
used to check the resistance of terminators.
You can use a DVM to help you find a break or a short in a network
cable.
DVMs are usually inexpensive battery operated devices that have
either a digital or needle read out, and two metal prongs attached to
the DVM by some wires a foot or more in length. By sending a
small current through the wires and out through the metal prongs,
resistance and voltages of terminators and wires can be measured.
Time-Domain Reflectometers
A Time-Domain Reflectometer (TDM) sends sound waves along a
cable and look for imperfections that might be caused by a break or
a short in the line. A good TDR will often be able to detect faults
on a cable to within a few feet.
Oscilloscope
An oscilloscope measures fluctuations in signal voltage and can
help find faulty or damaged cabling. Oscilloscopes are often more
expensive electronic devices that show the signal fluctuations on a
monitor.
NOTE
Digital Volt Meter (DVM)
Meters Voltmeters that test voltage
and resistance sometimes are called
Multimeters. There are also devices
called Ohmmeters used to test resistance only.
19.919-x CH12.i
426
8/28/98 2:07 PM
C h a p te r 1 2
Page 426
TRO UBL E S H O OTI NG
Other Tools
Several diagnostic software tools provide information on virtually
any type of network hardware, as well. A considerable number of
diagnostic software packages are available at a variety of prices.
A common software tool distributed with most network cards is a
Send/Receive package. This software tool allows two computers,
with network cards and cables, to connect to each other. This tool
does not rely on a networked operating system, nor can it be used to
send data. It will simply send packets from one computer to the
other, establishing that the network cards and underlying transmission media are connected and configured properly.
TROUBLESHOOTING TRANSMISSION
MEDIA AND OTHER NETWORK
COMPONENTS
Identify common errors associated with components required for
communications
Diagnose and resolve common connectivity problems with cards,
cables, and related hardware
Most network problems occur on the transmission media or with the
components that attach devices to the transmission media. All of
these components operate at the Physical, Datalink, and Network
levels of the OSI model. The components that connect PCs and
enable them to communicate are susceptible to many kinds of problems. The following sections discuss these important connectivity
and communication components and some of the problems associated with them:
á Cables and connectors
á Network adapter cards
á Modems
á Hubs and MSAUs
á Repeaters, Bridges, Gateways, and Routers
19.919-x CH12.i
8/28/98 2:07 PM
Page 427
C hapter 12
These components were introduced in previous chapters of this
book. (See Chapter 3, “Transmission Media,” Chapter 5, “Network
Adapter Cards,” and Chapter 6, “Connectivity Devices and Transfer
Mechanisms”) This chapter concentrates on troubleshooting guidelines.
Troubleshooting Cables and
Connectors
Most network problems occur at the OSI Physical layer, and cabling
is one of the most common causes. A cable might have a short or a
break, or it might be attached to a faulty connector. Tools such as
DVMs and TDRs help search out cabling problems.
Cabling problems can cause one of three major problems:
á An individual computer cannot access the network
á A group of computers cannot access the network
á None of the computers can access the network
When networks are configured using a star topology, an individual
cable break between the computer and hub or MASU will cause a
failure in communication between that individual computer and the
rest of the network. This type of cable break will not cause problems
for all of the other computers on the network.
A break in cables connecting multiple hubs together will cause a
communications outage between the computers on one side of the
cable and the computers on the other side of the cable break. In
most cases the communications between computers within the broken segment can continue.
In the case of a MSAU, if a cable connecting MSAUs together is
broken, this will often cause all computers on the ring to fail,
because the ring is not complete. A break in the cable on a Bus
topology will also cause all computers on the network segment to be
unable to communicate with any other computers on the network.
When troubleshooting any network, begin with the more obvious
physical problems. Make sure that all connectors are tight and
properly connected, that ground wires and terminators with the
proper resistance are used when required, and that manufacturer’s
TROU BLESHOOT IN G
427
19.919-x CH12.i
428
8/28/98 2:07 PM
C h a p te r 1 2
Page 428
TRO UBL E S H O OTI NG
specifications (such as cable grade, cable lengths, and maximum
number of nodes) are met and are consistent with the specifications
for the transmission medium.
Try the following checks when troubleshooting network cabling
problems:
á With 10BASE-T, make sure the cable used has the correct
number of twists to meet the data-grade specifications.
á Look for electrical interference, which can be caused by tying
the network cable together with monitor and power cords.
Fluorescent lights, electric motors, and other electrical devices
can cause interference if they are located too close to cables.
These problems can often be alleviated by placing the cable
away from these devices that generate electromagnetic interference or by upgrading the cable to one that has better shielding.
á Make sure that connectors are pinned properly and crimped
tightly.
á If excess shielding on coaxial cable is exposed, make sure it
doesn’t ground out the connector.
á Ensure that coaxial cables are not coiled tightly together. This
can generate a magnetic field around the cable, causing electromagnetic interference.
á On coaxial Ethernet LANs, look for missing terminators or
terminators with improper resistance ratings or resistance readings.
á Watch out for malfunctioning transceivers, concentrators, or
T-connectors. All of these components can be checked by
replacing the suspect devices.
á Test the continuity of the cable by using the various physical
testing devices discussed in the previous section, or by using a
software-based cable testing utility.
á Make sure that all the component cables in a segment are
connected. A user who moves his client and removes the
T-connector incorrectly can cause a broken segment.
á Examine cable connectors for bent or broken pins.
19.919-x CH12.i
8/28/98 2:07 PM
Page 429
C hapter 12
á On Token Ring networks, inspect the attachment of patch
cables and adapter cables. Remember, patch cables connect
MSAUs, and adapter cables connect the network adapter to
the MSAU.
One advantage of a Token Ring network is its built-in capability to
monitor itself. Token Ring networks provide electronic troubleshooting and, when possible, actually make repairs. When the
Token Ring network can’t make its own repairs, a process called beaconing narrows down the portion of the ring in which the problem
is most likely to exist. (See Chapter 4, “Network Topologies and
Architectures,” for more information on beaconing.)
Troubleshooting Network Adapter
Cards
Network problems often result from malfunctioning network
adapter cards. The process of troubleshooting the network adapter
works like any other kind of troubleshooting process: Start with the
simple. The following list details some aspects you can check if you
think your network adapter card might be malfunctioning:
á Make sure the cable is properly connected to the card.
á Confirm that you have the correct network adapter card driver
and that the driver is installed properly (see Chapter 5). Be
sure the card is properly bound to the appropriate transport
protocol (see Chapter 7, “Transport Protocols”).
á Make sure the network adapter card and the network adapter
card driver are compatible with your operating system. If you
use Windows NT, consult the Windows NT hardware compatibility list. If you use Windows 95 or another operating system, rely on the adapter card vendor specifications.
á Test for resource conflicts. Make sure another device isn’t
attempting to use the same resources (see Chapter 5). If you
think a resource conflict might be the problem, but you can’t
pinpoint the conflict using Windows NT Diagnostics,
Windows 95’s Device Manager, or some other diagnostic program, try removing all the cards except the network adapter
TROU BLESHOOT IN G
429
19.919-x CH12.i
430
8/28/98 2:07 PM
C h a p te r 1 2
Page 430
TRO UBL E S H O OTI NG
and then replacing the cards one by one. Check the network
with each addition to determine which device is causing the
conflict.
á Run the network adapter card’s diagnostic software. This will
often indicate which resource on the network card is failing.
á Examine the jumper and DIP switch settings on the card.
Make sure the resource settings are consistent with the settings
configured through the operating system.
á Make sure the card is inserted properly in the slot. Remove
and reseat the card.
á If necessary, remove the card and clean the connector fingers
(don’t use an eraser because it leaves grit on the card).
á Replace the card with one that you know works. If the connec-
tion works with a different card, you know the card is the
problem.
Token Ring network adapters with failure rates that exceed a preset
tolerance level might actually remove themselves from the network.
Try replacing the card. Some Token Ring networks also can experience problems if a Token Ring card set at a ring speed of 16 Mbps is
inserted into a ring using a 4 Mbps ring speed and vise versa.
Broadcast storms (discussed later in this chapter) are often caused by
faulty network adapters as well.
Troubleshooting HUBs and MSAUs
If you experience problems with a hub-based LAN, such as a
10BASE-T network, you often can isolate the problem by disconnecting the attached workstations one at a time. If removing one of
the workstations eliminates the problem, the trouble may be caused
by that workstation or its associated cable length. If removing each
of the workstations doesn’t solve the problem, the fault may lie with
the hub. Check the easy components first, such as ports, switches,
and connectors. Then use a different hub (if you have it) and see
whether the problem persists. If your hub doesn’t work properly, call
the manufacturer.
19.919-x CH12.i
8/28/98 2:07 PM
Page 431
C hapter 12
If you’re troubleshooting a Token Ring network, make sure the
cables are connected properly to the MSAUs, with ring-out ports
connecting to the ring-in ports throughout the ring. If you suspect
the MSAU, isolate it by changing the ring-in and ring-out cables to
bypass the MSAU. If the ring is now functional again, consider
replacing the MSAU. In addition, you might find that if your network has MSAUs from more than one manufacturer, they may not
be wholly compatible. Impedance and other electrical characteristics
can show slight differences between manufacturers, causing intermittent network problems. Some MSAUs (other than the 8228) are
active and require a power supply. These MSAUs fail if they have a
blown fuse or a bad power source. Your problem also might result
from a misconfigured MSAU port. MSAU ports using the hermaphrodite connector need to be reinitialized with the setup tool.
Removing drop cables and reinitializing each MSAU port is a quick
fix that is useful on relatively small Token Ring networks.
Isolating problems with patch cables, adapter cables, and MSAUs is
easier to do if you have a current log of your network’s physical
design. After you narrow down the problem, you can isolate potential problem areas from the rest of the network and then use a cable
tester to find the actual problem.
Troubleshooting Modems
A modem presents all the potential problems you find with any
other device. You must make sure that the modem is properly
installed, that the driver is properly installed, and that the resource
settings do not conflict with other devices. Modems also pose some
unique problems because they must connect directly to the phone
system, they operate using analog communications, and they must
make a point-to-point connection with a remote machine.
The online help files for both Windows NT and Windows 95
include a topic called the Modem Troubleshooter (see Figure 12.1).
The Modem Troubleshooter leads you to possible solutions for a
modem problem by asking questions about the symptoms. As you
answer the questions (by clicking the gray box beside your answer),
the Modem Troubleshooter zeroes in on more specific questions
until (ideally) it leads you to a solution. See Exercise 12.1 at the end
of this chapter for more on the Modem Troubleshooter.
TROU BLESHOOT IN G
431
19.919-x CH12.i
432
8/28/98 2:07 PM
C h a p te r 1 2
Page 432
TRO UBL E S H O OTI NG
Some common modem problems (in addition to the basic device
problems discussed earlier in this chapter, such as connectivity and
resource settings) are as follows:
á Dialing problems. The dialing feature is improperly configured.
For instance, the modem isn’t dialing 9 to bypass your office
switchboard, or it is dialing 9 when you’re away from your
office. The computer also could be dialing an area code or an
international code when it shouldn’t. Check the dialing properties for the connection.
á Connection problems. You cannot connect to another modem.
FIGURE 12.1
Windows NT Help showing topics on troubleshooting modems.
Your modem and the other modem might be operating at different speeds. Verify that the maximum speed setting for your
modem is the highest speed that both your modem and the
other modem can use. Also make sure the Data Bits, Parity,
and Stop Bits settings are consistent with the remote computer.
á Digital Phone Systems. You cannot plug a modem into a tele-
phone line designed for use with digital phone systems. These
digital phone systems are commonplace in most office environments.
á Protocol problems. The communicating devices are using
incompatible line protocols. Verify that the devices are configured for the same or compatible protocols. If one computer
initiates a connection using PPP, the other computer must be
capable of using PPP.
Repeaters, Bridges, and Routers
Issues dealing with repeaters, bridges and routers are often more
technically advanced, than those covered in a book such as
Networking Essentials. Companies such as Cisco, Bay Networks, and
3Com have their own dedicated books and courses on dealing with
the installation, configuration and troubleshooting of repeaters,
bridges and routers. In general, there are some basic troubleshooting
steps you can do when working with these three devices.
As mentioned in Chapter 2, “Networking Standards,” and Chapter
6, repeaters are responsible for regenerating a signal sent down the
transmission media. Problems with repeaters are that they do not
19.919-x CH12.i
8/28/98 2:07 PM
Page 433
C hapter 12
work—that is, the signal is not being regenerated. If this is the case,
the signal being sent to devices on the other side of the repeater
from the sending device will not receive the signal.
Problems with bridges are almost identical to that of a repeater. The
signal being sent to devices on the other side of the bridge from the
sending device will not receive the signal. Other issues with bridges
are that the table of what devices are on what interface of the bridge
can get corrupt. This can lead from one to all machines not receiving packets on the network. Diagnostic utilities provided by the
bridge’s manufacturer can resolve this type of problem.
Problems with routers can be complex, and troubleshooting them
often involves a high level of understanding of the different protocols in use on the network, as well as the software and commands
used to program a router. There are generally two types of router
problems.
The first router problem that is commonly found is that packets are
just not being passed through, because the router is “dead,” or simply not functioning. The second common problem with routers is
that the routing tables, within the routers, are corrupted or incorrectly programmed. This problem will either lead to computers on
different networks being unable to communicate with each other or
to the fact that certain protocols simply do not work.
HANDLING BROADCAST STORMS
Resolve broadcast storms.
A broadcast storm is a sudden flood of broadcast messages that clogs
the transmission medium, approaching use of 100 percent of the
bandwidth. Broadcast storms cause performance to decline and, in
the worst case, computers cannot even access the network. The cause
of a broadcast storm is often a malfunctioning network adapter, but
a broadcast storm also can be caused when a device on the network
attempts to contact another device that either doesn’t exist or for
some reason doesn’t respond to the broadcast.
If the broadcast messages are viable packets (or even error-filled but
partially legible packets), a network-monitoring or protocol-analysis
tool often can determine the source of the storm (see Chapter 11). If
the broadcast storm is caused by a malfunctioning adapter throwing
TROU BLESHOOT IN G
433
19.919-x CH12.i
434
8/28/98 2:07 PM
C h a p te r 1 2
Page 434
TRO UBL E S H O OTI NG
illegible packets onto the line, and a protocol analyzer can’t find the
source, try to isolate the offending PC by removing computers from
the network one at a time until the line returns to normal. (For
more information, see “Troubleshooting Network Adapter Cards,”
earlier in this chapter.)
TROUBLESHOOTING PROTOCOLS
When discussing troubleshooting protocols, three protocol stacks are
at issue: NetBEUI, NWLink, and TCP/IP. It is these three protocol
stacks that can be used by Microsoft networks to communicate over
the network. It is important that computers wishing to communicate with one another run the same transport protocol, or else communication will not occur.
NetBEUI
NetBEUI is a relatively simple protocol to troubleshoot. The most
important thing to remember is that NetBEUI is a non-routable
protocol, and will not pass through routers. The other important
issue concerning NetBEUI is that the computers address each other
by their NetBIOS names. Thus each NetBIOS name must be unique
(this holds true for all protocols on computers utilizing Microsoft
operating systems). This latter issue is detected by Microsoft operating systems (DOS, Windows 95, Windows NT), when a computer
first initializes on the network.
NWLink (IPX/SPX)
NWLink (IPX/SPX) will generate communication problems if the
incorrect frame type is chosen. Two computers running different
frame types will not communicate with one another. A computer
can run more than one frame type, and this is often the case when
running an Ethernet network, as an Ethernet network has the greatest selection of frame types. Frame types are discussed in detail in
Chapter 2.
Older Novell systems (systems prior to NetWare 3.12) had a default
frame type of 802.3. All new systems use a frame type of 802.2.
Both systems can use both frame types. Windows 95 and Windows
19.919-x CH12.i
8/28/98 2:07 PM
Page 435
C hapter 12
NT will install the 802.2 frame type if it is detected but not 802.3.
If your network is running both the 802.2 and 802.3 frame types,
you will need to manually add the 802.3 frame type to your
Windows 95 and Windows NT computers.
TCP/IP
TCP/IP relies on a strong knowledge of the protocol stack in order
to troubleshoot it. Basic configuration of the protocol includes the
correct IP address information being set on the computer. If you
wish to communicate to other network segments, then a proper
default gateway will need to be set.
Other issues involving TCP/IP are that services are running
correctly. If you are using a DHCP server to hand out IP addressing
information to the computers, make sure that the DHCP server is
running correctly. Also, make sure that the server has not run out of
addresses, or is not giving out addresses that exist on another DHCP
computer.
If name-to-address resolution is at issue, make sure that the DNS
server (used for non-Microsoft computers) and the WINS server
(used to locate Microsoft NetBIOS names) are both running, or
contain the name to IP address resolution being requested.
TROUBLESHOOTING NETWORK
PERFORMANCE
Identify and resolve network performance problems.
If your network runs slower than it used to run (or slower than it
ought to run), the problem might be that the present network traffic
exceeds the level at which the network can operate efficiently. Some
possible causes for increased traffic are new hardware (such as a new
workstation) or new software (such as a network computer game or
some other network application). A generator or another mechanical
device operating near the network could cause a degradation of network performance. In addition, a malfunctioning network device
could act as a bottleneck. Ask yourself what has changed since the
last time the network operated efficiently, and begin there with your
troubleshooting efforts.
TROU BLESHOOT IN G
435
19.919-x CH12.i
436
8/28/98 2:07 PM
C h a p te r 1 2
Page 436
TRO UBL E S H O OTI NG
Some of the techniques described in previous chapters can help you
troubleshoot network performance. A performance monitoring tool,
such as Windows NT’s Performance Monitor, can help you look for
bottlenecks that are adversely affecting your network. See Chapter
11 for more information on Performance Monitor.
The monitoring and record-keeping procedures discussed in Chapter
11 also can help you troubleshoot network performance by providing you with baseline performance data that you can use to gauge
later fluctuations.
For instance, the increased traffic could be the result of increased
usage. If usage exceeds the capacity of the network, you might want
to consider expanding or redesigning your network. You also might
want to divide the network into smaller segments by using a router
or a bridge to reduce network traffic. A protocol analyzer can help
you measure and monitor the traffic at various points on your network.
HANDLING OTHER NETWORK
PROBLEMS
The following list details some other common problems that could
affect your network:
á Operating system conflicts. Operating system upgrades some-
times can cause older programs to become incompatible with
the operating system itself. This problem is compounded in
network environments because, during the transition to a new
network operating system, some servers will run the new version for a period of time while others are still running the previous version. Microsoft recommends that you perform a test
upgrade on an isolated part of the network to ensure that all
hardware and software systems function properly when the
upgrade is made.
á Server crashes. A server disk crash can be disastrous if you aren’t
adequately prepared for it. You should devise a system of regular backups and, depending on the nature of your data, explore
other safeguards such as a RAID fault tolerant system. (Refer
to Chapter 9, “Disaster Recovery.”)
19.919-x CH12.i
8/28/98 2:07 PM
Page 437
C hapter 12
á Power fluctuations. A small fluctuation in the power supply can
make the network misbehave. If the power goes off completely—even for a moment—the whole network could shut down,
causing users to lose their work in progress. A disorderly shutdown also can cause problems with file servers. The best solution is to prepare for a power outage before it happens.
Connect each server to an Uninterruptible Power Supply
(UPS), and encourage your users to perform occasional saves
as they work.
If you implement all the measures discussed so far and you still
experience problems, your next step may be to consult the experts.
Or, even before you start your own troubleshooting, you may want
to consult the available information to learn more about the problem. The next section discusses some online and offline sources of
help.
GETTING SUPPORT
An important aspect of troubleshooting is knowing where to turn
for critical information on your network environment. Many online
and offline sources can provide troubleshooting information. Some
of these sources (in addition to the online help provided with your
operating system), include the following:
á Vendor documentation and help lines. Hardware and software
vendors often provide troubleshooting tips with the owner’s
documentation. Vendors also often provide technical assistance
by phone.
á Bulletin board services. A number of electronic bulletin boards
supply networking information. See vendor documentation for
more information on how to reach a particular vendor’s official
BBS.
á The Internet. The major network vendors all sponsor active
forums and newsgroups on the Internet, CompuServe, and
other online services. Most of this information is also supplied
through the Web. Updated drivers, white papers, technical
specifications, and knowledgebases can often be found at vendors’ Web sites. Also refer to your vendors’ documentation.
TROU BLESHOOT IN G
437
19.919-x CH12.i
438
8/28/98 2:07 PM
C h a p te r 1 2
Page 438
TRO UBL E S H O OTI NG
á CD-ROMs. Several vendors now market CD-ROMs with net-
work and PC hardware information. Windows NT Server’s
Books Online (located on the Windows NT Installation CDROM) provides an additional layer of documentation that isn’t
found with online help. Microsoft’s TechNet contains product
information, technical information, articles, and announcements. TechNet is available on a subscription basis through
Microsoft (call 800-344-2121). Novell’s NSEPro CD-ROM is
a NetWare oriented encyclopedia of network information. The
Micro House Technical Library (MHTL) is another impressive
database of technical information. The MHTL addresses such
items as BIOS settings for IDE drives and jumper settings for
popular peripheral boards. The MHTL comes with a rich collection of informative illustrations.
C A S E S T U DY : T R O U B L E S H O O T I N G
A
NETWORK
ESSENCE OF THE CASE
SCENARIO
The essence of the case is as follows:
You are brought into XYZ company to help troubleshoot the network. The network is a small 60user Ethernet network using a bus topology running Thinnet wire. There is a router placed in the
middle of this bus topology, essentially dividing
the LAN into two equally sized segments of 30
computers each.
• The LAN runs a bus topology.
• There are two networked segments connected by a router.
• Network connectivity is not available for
one segment.
• The computers on the affected segment
can for a short time communicate.
At issue is that at certain times, none of the
computers are able to communicate with each
other on one of the network segments. The other
network segment is able to communicate only
with PCs on its own segment.
As noted, this situation is not always the case. At
certain times, the network is fully functional. At a
given point in the day the network becomes inoperable. When everyone on the affected segment
shuts down their PCs and turns them back on
again, the network is operable for a few minutes,
but then returns to an inoperable state.
19.919-x CH12.i
8/28/98 2:07 PM
Page 439
C hapter 12
TROU BLESHOOT IN G
C A S E S T U DY : T R O U B L E S H O O T I N G
A
A N A LY S I S
communicate with one another even for a short
period after they were shut down.
This case study is a nice example of a broadcast
storm. As discussed earlier in the chapter, a
broadcast storm is caused by a “chatty” or malfunctioning network card. This chatty card causes
broadcast packets, or packets addressed to all
devices on the network. The broadcast storm can
saturate the bandwidth on a network and cause
all of the computers to not be able to communicate with each other.
The problems experienced by the network in this
case study are also commonly attributed to broken cables, terminators, or a malfunctioning
router.
A broken Thinnet cable would have permanently
affected all computers on the one network segment. The fact that the computers could transmit
to one another after they were all shut down and
restarted, negates the possibility of a broken
cable being suspect.
A missing terminator would also cause all PCs on
the network segment to not be able to communicate with each other. Just as with the issue with
the broken cable, the PCs would not be able to
439
NETWORK
If the router were the problem, two things would
happen. The first would be an absence of communication between the two network segments.
Second, only certain computers would be able to
communicate between the network segments. A
faulty router would not cause one network segment to be unable to communicate with another.
The compelling reason for the broadcast storm
being the culprit is the fact that the computers
can communicate with each other once they are
all rebooted. Shortly after they reboot, the network becomes inaccessible. This inaccessibility
is due to the fact that the computer with the
chatty network card has come on line.
You can track down the computer with the chatty
network card in either of two ways. Run a protocol analyzer to see if the broadcasts are issuing
a source address, and cross reference this
source address with your documentation. Another
method would be to systematically shut down
each PC, one at a time, until the faulty network
card is found.
19.919-x CH12.i
440
8/28/98 2:07 PM
C h a p te r 1 2
Page 440
TRO UBL E S H O OTI NG
CHAPTER SUMMARY
KEY TERMS
• Protocol analyzers
• Digital voltmeters
• Time-Domain Reflectometer
(TDM)
• Oscilloscope
• Broadcast storm
• Network adapter card
• Modem
Troubleshooting is an essential part of network operations. The best
kind of troubleshooting is, of course, anticipating problems before
they occur; however, in spite of all your efforts, you’ll eventually
need to track down a problem that is stopping or slowing your network. This chapter looked at general troubleshooting strategies and
at solutions for each of the problem areas identified by Microsoft in
the Networking Essentials test objectives, as follows:
á Problems with communication components
á Connectivity problems
á Broadcast storms
á Network performance problems
• Hub
• MSAU
• Repeater
• Bridge
• Router
• NetBEUI
• NWLink
• TCP/IP
This chapter also looked at online and offline sources of troubleshooting information.
19.919-x CH12.i
8/28/98 2:07 PM
Page 441
C hapter 12
TROU BLESHOOT IN G
441
A P P LY Y O U R L E A R N I N G
Exercises
1. Click the Start menu and choose Programs/Books
Online.
12.1 Modem Troubleshooter
2. Windows NT prompts you for the location of
the Books Online files. The files are located on
the Windows NT Installation CD-ROM in the
Support directory’s Books subdirectory.
Objective: Learn how to access Windows NT’s or
Windows 95’s Modem Troubleshooter. This exercise
addresses Microsoft’s exam objective: Identify common
problems associated with components required for
communications.
Estimated time: 10 minutes
Modem Troubleshooter is part of Windows NT’s
online help system. The easiest way to access it is to
start Help and search for modems in the index.
1. Click the Start button and choose Help.
2. In the Help Topics dialog box, click the Index
tab. Enter modem in the search box at the top of
the screen.
3. In the Books Online main dialog box, click the
Contents tab. Double-click the book icon to
reveal the major subcategories for Books Online
(see Figure 12.2).
4. Browse through the topics in Books Online.
Notice the extensive Networking Supplement
section devoted to networking issues. Try to get a
feeling for the kinds of questions that are best
answered by Books Online.
5. When you are finished, close the Books Online
dialog box.
3. Look for the troubleshooting subtopic under the
modems topic in the index. Double-click
Troubleshooting. The Modem Troubleshooter
will appear (refer to Figure 12.1).
4. Browse through the Modem Troubleshooter’s
topics. Click the gray box to the left of each
symptom for a look at possible causes or more
diagnostic questions.
5. When you’re finished, close the Help window.
12.2 Windows NT Books Online
Objective: Access Windows NT’s Books Online, a
CD-ROM-based source of configuration and troubleshooting information.
Estimated time: 10 minutes
FIGURE 12.2
Windows NT’s Books Online contains information not found
in online help.
19.919-x CH12.i
442
8/28/98 2:07 PM
C h a p te r 1 2
Page 442
TRO UBL E S H O OTI NG
A P P LY Y O U R L E A R N I N G
Review Questions
1. What are four tools that can be used to troubleshoot network problems? Briefly define each.
C. TDR
D. MSDL
4. Most network problems occur at what OSI layer?
2. What will a broken or improperly terminated
cable in a BUS topology cause?
A. Physical
3. What is a broadcast storm and what causes it?
B. Data Link
4. What Microsoft utility can be used to detect bottlenecks within a computer? What Microsoft program can be used to detect bottlenecks on the
network?
C. Network
D. Session
5. What is a sudden, unexpected flood of broadcast
messages on the network is known as?
A. Net frenzy
Exam Questions
1. Which three of the following are troubleshooting
steps in Microsoft’s five-step troubleshooting
process?
A. Collect information to identify the symptoms.
B. Develop a list of possible causes.
B. Tornado
C. Broadcast storm
D. Electric shower
6. What device sends sound waves down the cable
to look for imperfections?
C. Reboot the server.
A. DVMs
D. Set the problem’s priority.
B. Oscilloscopes
C. TDRs
2. What does MSDL stands for?
D. None of the above
A. Minor Switching Delay Log
B. Microsoft Storage Device Language
C. Microsoft Domain License
D. Microsoft Download Library
3. What can you use to look for breaks in network
cables by measuring cable voltage?
7. Which of the following would degrade network
performance? Choose three.
A. A generator or electrical device near the network
B. A networked computer game
A. Protocol analyzer
C. A sudden, disorderly shutdown of a workstation
B. DVM
D. New hardware
19.919-x CH12.i
8/28/98 2:07 PM
Page 443
C hapter 12
TROU BLESHOOT IN G
443
A P P LY Y O U R L E A R N I N G
8. Which two of the following are possible problems
with Token Ring network adapters?
A. Broadcast messages from the card are not
timed right.
B. The card is not bound to a network service.
C. The card removed itself from the network.
D. A 16Mbps card is placed on a 4Mbps ring.
Required result: You need to see which systems
are generating the most traffic on the network
cable segments.
Optional result 1: You wish to see which
resources are being used on the workstation.
Optional result 2: You wish to log the information about system resource usage for analysis at a
later time.
9. You want to have a set of tools to be able to isolate problems with your transmission media. The
object is to have the tools to analyze and troubleshoot problems on your Thinnet Ethernet network.
Suggested Solution: You load Performance
Monitor on the computers you wish to analyze
and select the objects you wish to monitor.
Required Result: You need to be able to test for
breaks in the cables on the network
B. This solution will obtain the required result
and one of the optional results.
Optional result 1: You wish to be able to test for
the resistance of terminators attached to the ends
of your cables.
C. This solution will obtain the required result.
A. This solution will obtain the required result
and both optional results.
D. This solution does not satisfy the required
result.
Optional result 2: You wish to be able to track
network traffic on your cables.
Suggested solution: You purchase an oscilloscope
and a Time Domain Reflectometer.
A. This solution will obtain the required result
and both optional results.
B. This solution will obtain the required result
and one of optional results.
C. This solution will obtain the required result.
D. This solution does not satisfy the required
result.
10. You wish to analyze the performance of your network. You would like to see what systems are
generating the most network traffic.
Answers to Review Questions
1. Four tools can be used to troubleshoot a network:
• Protocol analyzer. This hardware or software
tool can capture and analyze traffic and packets on a network. The Microsoft software version of this tool is called Network Monitor.
• Digital Volt Meter (DVM). This electronic
measuring device can be used to check the
voltage of cable segments, allowing you to see
if there are breaks on a cable, and the resistance of terminators on cable segments.
• Time-Domain Reflectometer (TDR). A TDR
sends sound waves along a cable to look for
19.919-x CH12.i
444
8/28/98 2:07 PM
C h a p te r 1 2
Page 444
TRO UBL E S H O OTI NG
A P P LY Y O U R L E A R N I N G
imperfections that may be caused by a break
or short in the line.
• Oscilloscope. This device measures the fluctuation in signal voltage to help find faulty or
damaged cables.
See the section titled “Using Troubleshooting
Tools.”
2. On a BUS topology, a broken cable segment or
an improperly terminated cable segment will
cause all devices on that cable segment to not be
able to communicate with one another. See the
section titled “Troubleshooting Cables and
Connectors.”
3. A broadcast storm is a network segment that is
being saturated by broadcast packets. Packets that
are broadcast are intended for all recipients. A
broadcast storm will cause network traffic to slow
or even halt traffic on the network segment.
A faulty or “chatty” network card is often the
cause of a broadcast storm. See the section titled
“Handling Broadcast Storms.”
4. Microsoft’s Performance Monitor is used to track
objects and counters within a computer.
Microsoft’s Network Monitor is designed to track
packets and traffic outside a computer. See the
section titled “Troubleshooting Network
Performance.”
Answers to Exam Questions
1. A, B, D. Rebooting a server is a task in troubleshooting problems, but is not part of
Microsoft’s five-step troubleshooting process.
Many problems on a network do not involve a
server, thus rebooting a server will be of no avail.
See the section titled “Initiating the
Troubleshooting Process.”
2. D. MSDL is the official acronym for the
Microsoft Download Library. See the section
titled “Getting Support.”
3. B. A Digital Voltmeter uses electrical signals to
look for breaks in a cable. A TDR uses sound signals. MSDL is a reference tool for information. A
protocol analyzer analyzes packets on a network.
See the section titled “Using Troubleshooting
Tools.”
4. A. Most network problems occur at the OSI
Physical layer, or in other words, involve the
transmission media. See the section titled
“Establishing Troubleshooting Connectivity and
Communication.”
5. C. A broadcast storm is the official definition of a
sudden broadcast of network messages. See the
section titled “Handling Broadcast Storms.”
6. C. Time Domain Reflectometers use sound waves
to test for imperfections. DVMs use electrical signals. Oscilloscopes do not look for cable imperfections, but instead measure electrical signals.
See the section titled “Using Troubleshooting
Tools.”
7. A, B, D. C is not correct because the shutting
down of a PC on the network causes one fewer
PC to generate traffic on the network. See the
section titled “Troubleshooting Network
Performance.”
8. C, D. A is incorrect because Token Ring networks can transfer information on the transmission media only when they have the token. B is
incorrect because cards do not bind to services,
but to transport protocols. See the section titled
“Troubleshooting Network Adapter Cards.”
19.919-x CH12.i
8/28/98 2:07 PM
Page 445
C hapter 12
TROU BLESHOOT IN G
445
A P P LY Y O U R L E A R N I N G
9. C. A TDR will test for cable breaks, but does not
test for the resistance of a terminator. This is
done by a Digital Volt Meter. An oscilloscope
checks the flow of a current. Neither a TDR or
an oscilloscope will analyze network traffic.
10. D. Performance Monitor will do both of the
optional results but will not achieve the required
result, because Performance Monitor does not
trace traffic on the network transmission media.
To analyze network traffic on the transmission
media, you would need to use a protocol analyzer
such as Network Monitor.
Suggested Readings and Resources
For all Web sites, do searches on the key words of
“troubleshooting” and the name of the component
you wish to troubleshoot.
1. Sirockman, Jason. MCSE Training Guide:
Windows NT 4 Enterprise, 2nd Edition. New
Riders, 1998.
2. Microsoft Technet CD. For more information,
go to www.microsoft.com/technet
3.
www.microsoft.com/support
4.
www.cisco.com
5.
www.novell.com
6. Brooks, Charles, and Marcraft International.
A+ Certification Training Guide. New Riders,
1998.
7. Mueller, Scott. Upgrading and Repairing PCs,
Eighth Edition. Que, 1997.
19.919-x CH12.i
8/28/98 2:07 PM
Page 446
20.919-x Part5
8/28/98 2:13 PM
Page 447
P A R T
FINAL REVIEW
Fast Facts: Networking Essentials
Fast Facts: Windows NT Server 4
Fast Facts: Windows NT Server 4 Enterprise
Fast Facts: Windows NT Workstation 4
Study and Exam Prep Tips
Practice Exam
V
20.919-x Part5
8/28/98 2:13 PM
Page 2
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 449
Twelve chapters of this book have looked at objectives and components of the Microsoft Networking
Essentials exam. After reading all of that, what is it
that you must really know? What should you read
as you sit and wait in the parking lot of the testing
center—right up until the hour before going in to
gamble your $100 and pride?
The following material covers the salient points of
the 12 previous chapters and the points that make
excellent test fodder. Although there is no substitute for real-world, hands-on experience, knowing
what to expect on the exam can be equally meaningful. The information that follows is the networking equivalent of Cliffs Notes, providing the information you must know in each of the four sections to
pass the exam. Don’t just memorize the concepts
given; attempt to understand the reason why they
are so, and you will have no difficulties passing the
exam.
STANDARDS AND
TERMINOLOGY
The Standards and Terminology section is designed to
test your understanding and knowledge of terms used
in networking, as well as some of the more common
standards that have been implemented in the industry.
Define Common Networking
Terms for LANs and WANs
The Networking Essentials exam does not really test on
definitions of terms. You are asked questions though,
and, based on these questions, you need to understand
the definitions of the terms used in order to successfully answer the questions.
Fast Facts:
Networking
Essentials
Exam
15 918-1 FastFacts.i Network
450
9/1/98 6:58 AM
Page 450
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
The best mechanism to study for this area would be to
be able to review the key terms found in every chapter
and provide the correct definition for each term. Below
is a list of some of the more general networking terms
you should be aware of.
á peer-to-peer networking. A networking model
where both the services and the client are performed by the same computer.
á client/server networking. A networking model
where a specific role of providing services or acting as a client (not both) is performed by a computer.
á centralized computing. A form of computing
where all the processing is done by one central
computer.
á distributed computing. A form of computing
where all the processing is shared by many different computers.
á ethernet network. This type of a network is run
as a logical bus, but can take on the physical
topology of a bus or a star. The concentrator used
by these computers, when in a star topology, is
called a hub. This type of network is known as a
contention-based network because each device
contends with every other device for network
access.
á LAN. Also known as a Local Area Network.
Often characterized by fast transmission speeds
and short distances between devices, and by the
fact that the company running the network has
control over all devices and transmission media.
á WAN. Also known as a Wide Area Network.
When compared to a LAN, a WAN is often characterized by lower data transmission rates and the
coverage of long distances, and by the fact that a
third party is involved with the supply and maintenance of the transmission media.
á file services. Services allowing for the storage and
access of files.
á print services. Services that allow the sharing of a
printer.
á file and print server. A server that provides file
and print services.
á application server. A server that provides some
high-end application used by many different
computers.
á token-ring network. A network that follows a
logical topology of a ring, but a physical topology
of a star. The computers are connected to a concentrator known as an MSAU or MAU.
Computers rely on the possession of a token
before the transmission of data on the network.
This type of network is known as a deterministic
network.
Compare a File and Print
Server with an Application
Server
A file server is a service that is involved with giving
access to files and directories on the network. The purpose of the file server is to give large numbers of users
access to a centrally stored set of files and directories.
A print server is a computer or device that gives large
number of users access to a centrally maintained printing device. A computer that is a file server often acts as
print server, too. These types of computers are known
as file and print servers.
An application server is responsible for running applications such as Exchange Server or SQL Server on the
network. Application servers perform services that often
require a more advanced level of processing than a
user’s personal computer is able to provide.
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 451
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
Compare User-Level Security
with Access Permission
Assigned to a Shared Directory
on a Server
User-level security is a security model in which access
to resources is given on a user-by-user basis, a groupby-group basis, or both. This type of access restriction
allows an administrator to grant access to resources and
affords users seemless access to those resources. Userlevel security is offered by Windows NT in both the
workgroup and domain models.
The permissions to a shared directory are:
451
RAM, more hard drive space, and a faster CPU than
the other machines. A server services requests from
clients. These requests could be for the use of files and
printers, application services, communication services,
and database services.
Clients are the computers on which users work. These
computers typically are not as powerful as servers.
Client computers are designed to submit requests to
the server.
Peer-to-peer networks are made up of several computers
that play the roles of both a client and a server; thus
there is no dedicated computer running file and printer
services, application services, communication services,
or database services.
á Read. The user is allowed to read files within a
share. He can also see all files and subdirectories.
á Change. The user can modify existing files and
directories and create new files and directories
within the share.
á Full Control. The user can see, modify, delete,
and take ownership of all files and directories
within the share.
á No Access. The user cannot access any files or
directories within the share.
Share-level permissions apply to anyone accessing the
share over the network and do not apply to users who
are interactive on the computer where the share resides.
Share-level permissions can be set on both FAT and
NTFS partitions.
Compare a Client/Server
Network with a Peer-to-Peer
Network
A client/server network is one in which a computer has
a specific role. A server is a computer, often with more
Compare the Implications of
Using Connection-Oriented
Communications with
Connectionless
Communications
In general, connection-oriented communication differs
from connectionless communication as follows:
á Connection-oriented mode. Error correction
and flow control are provided at internal nodes
along the message path.
á Connectionless mode. Internal nodes along the
message path do not participate in error correction and flow control.
In connection-oriented mode, the chain of links
between the source and destination nodes forms a kind
of logical pathway connection. The nodes forwarding
the data packet can track which packet is part of which
connection. This enables the internal nodes to provide
flow control as the data moves along the path. For
example, if an internal node determines that a link is
15 918-1 FastFacts.i Network
452
9/1/98 6:58 AM
Page 452
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
malfunctioning, the node can send a notification
message backward through the path to the source computer. Furthermore, because the internal node distinguishes among individual, concurrent connections in
which it participates, this node can transmit (or forward) a “stop sending” message for one of its connections without stopping all communications through the
node. Another feature of connection-oriented communication is that internal nodes provide error correction
at each link in the chain. Therefore, if a node detects
an error, it asks the preceding node to retransmit.
SPX and TCP are two major examples of connectionoriented protocols.
Connectionless mode does not provide these elaborate
internal control mechanisms; instead, connectionless
mode relegates all error-correcting and retransmitting
processes to the source and destination nodes. The end
nodes acknowledge the receipt of packets and retransmit if necessary, but internal nodes do not participate
in flow control and error correction (other than simply
forwarding messages between the end nodes).
IPX and UDP are two major examples of connectionoriented protocols.
support dial-up access to networks based on the
Internet transport protocols. SLIP is a simple protocol
that functions at the Physical layer, whereas PPP is a
considerably enhanced protocol that provides Physical
layer and Data Link layer functionality.
Windows NT supports both SLIP and PPP from the
client end using the Dial-Up Networking application.
On the server end, Windows NT RAS (Remote Access
Service) supports PPP but doesn’t support SLIP. In
other words, Windows NT can act as a PPP server but
not as a SLIP server.
PPP
PPP was defined by the Internet Engineering Task
Force (IETF) to improve on SLIP by providing the following features:
á Security using password logon
á Simultaneous support for multiple protocols on
the same link
á Dynamic IP addressing
á Improved error control
The advantage of connectionless mode is that connectionless communications can be processed more quickly
and more simply because the internal nodes only forward data and thus don’t have to track connections or
provide retransmission or flow control.
Different PPP implementations might offer different
levels of service and negotiate service levels when connections are made. Because of its versatility, interoperability, and additional features, PPP has surpassed SLIP
as the most popular serial-line protocol.
Distinguish Whether SLIP or
PPP Is Used as the
Communications Protocol for
Various Situations
SLIP
Two other standards vital to network communication
are Serial Line Internet Protocol (SLIP) and Point-toPoint Protocol (PPP). SLIP and PPP were designed to
SLIP is most commonly used on older systems or for
dial-up connections to the Internet via SLIP-server
Internet hosts.
Developed to provide dial-up TCP/IP connections,
SLIP is an extremely rudimentary protocol that suffers
from a lack of rigid standardization in the industry,
which sometimes hinders different vendor implementations of SLIP from operating with each other.
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 453
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
Certain dial-up configurations cannot use SLIP for the
following reasons:
á SLIP supports the TCP/IP transport protocol
only. PPP, however, supports TCP/IP, as well as a
number of other transport protocols, such as
NetBEUI, IPX, AppleTalk, and DECnet. In
addition, PPP can support multiple protocols
over the same link.
á SLIP requires static IP addresses. Because SLIP
requires static, or preconfigured, IP addresses,
SLIP servers do not support the Dynamic Host
Configuration Protocol (DHCP), which assigns
IP addresses dynamically or when requested.
(DHCP enables clients to share IP addresses so
that a relatively small number of IP addresses can
serve a larger user base.) If the dial-up server uses
DHCP to assign an IP address to the client, the
dial-up connection won’t use SLIP.
453
á Network Interface Card (NIC). Operates at the
Data Link layer. A NIC is responsible for converting information in a computer to a signal that
will be sent on the transmission media.
á Bridge. Operates at the Data Link layer of the
OSI mode. A bridge is responsible for isolating
network traffic on a cable segment. It performs
this task by building address tables that contain
the MAC address or hardware addresses of
devices on ether side of it.
á Router. Operates at the Network layer of the
OSI model. It is responsible for connecting different segments that have dissimilar logical network addresses.
á Gateway. Can appear at any level of the OSI
model but is primarily seen at the Network layer
and higher. The purpose of a gateway is to convert one network protocol to another.
á SLIP does not support dynamic addressing
through DHCP so SLIP connections cannot
dynamically assign a WINS or DNS server.
Define the Communication
Devices that Communicate at
Each Level of the OSI Model
á Repeater. Operates at the Physical layer of the
OSI model. The purpose of a repeater is to regenerate a signal, allowing a signal to travel beyond
the maximum distance specified by the transmission media.
á Hub. Operates at the Physical layer. A hub is a
concentrator that connects 10BASE-T cabling
together on an Ethernet network. Some hubs also
have the capability to act as a repeater.
á MSAU. Operates at the Physical layer. An MSAU
performs the same purpose of a hub, but is used
on token-ring networks.
Describe the Characteristics
and Purpose of the Media Used
in IEEE 802.3 and IEEE 802.5
Standards
The various media types used by the IEEE 802.3 and
802.5 are discussed below.
IEEE 802.3
This standard defines characteristics related to the
MAC sublayer of the Data Link layer and the OSI
Physical layer. Except for one minor distinction—frame
type—IEEE 802.3 Ethernet functions identically to
DIX Ethernet v.2.
The MAC sublayer uses a type of contention access
called Carrier Sense Multiple Access with Collision
Detection (CSMA/CD). This technique reduces the incidence of collision by having each device listen to the
15 918-1 FastFacts.i Network
454
9/1/98 6:58 AM
Page 454
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
network to determine whether it’s quiet (“carrier sensing”); a device attempts to transmit only when the network is quiescent. This reduces but does not eliminate
collisions because signals take some time to propagate
through the network. As devices transmit, they continue to listen so they can detect a collision should it
occur. When a collision occurs, all devices cease transmitting and send a “jamming” signal that notifies all
stations of the collision. Each device then waits a random amount of time before attempting to transmit
again. This combination of safeguards significantly
reduces collisions on all but the busiest networks.
The IEEE 802.3 Physical layer definition describes signaling methods (both baseband and broadband), data
rates, media, and topologies. Several Physical layer variants also have been defined. Each variant is named following a convention that states the signaling rate (1 or
10) in Mbps, baseband (BASE) or broadband
(BROAD) mode, and a designation of the media characteristics.
The following list details the IEEE 802.3 variants of
transmission media:
á lBASE5. This 1-Mbps network utilizes UTP
cable with a signal range up to 500 meters (250
meters per segment). A star physical topology is
used.
á 10BASE5. Typically called Thick Ethernet, or
Thicknet, this variant uses a large diameter (10
mm) “thick” coaxial cable with a 50-ohm impedance. A data rate of 10 Mbps is supported with a
signaling range of 500 meters per cable segment
on a physical bus topology.
á 10BASE2. Similar to Thicknet, this variant uses
a thinner coaxial cable that can support cable
runs of 185 meters. (In this case, the “2” only
indicates an approximate cable range.) The transmission rate remains at 10 Mbps, and the physical topology is a bus. This variant typically is
called Thin Ethernet, or Thinnet.
á 10BASE-F. This variant uses fiber-optic cables to
support 10-Mbps signaling with a range of four
kilometers. Three subcategories include 10BASEFL (fiber link), 10BASE-FB (fiber backbone), and
10BASE-FP (fiber passive).
á 10BROAD36. This broadband standard supports
channel signal rates of 10 Mbps. A 75-ohm coaxial cable supports cable runs of 1,800 meters (up
to 3,600 meters in a dual-cable configuration)
using a physical bus topology.
á 10BASE-T. This variant uses UTP cable in a star
physical topology. The signaling rate remains at
10 Mbps, and devices can be up to 100 meters
from a wiring hub.
á 100BASE-X. This proposed standard is similar to
10BASE-T but supports 100 Mbps data rates.
IEEE 802.5
The IEEE 802.5 standard was derived from IBM’s
Token Ring network, which employs a ring logical
topology and token-based media-access control. Data
rates of 1, 4, and 16 Mbps have been defined for this
standard.
Explain the Purpose of NDIS
and Novell ODI Network
Standards
The Network Driver Interface Specification (NDIS),
a standard developed by Microsoft and the 3Com
Corporation, describes the interface between the
network transport protocol and the Data Link layer
network adapter driver. The following list details the
goals of NDIS:
á To provide a vendor-neutral boundary between
the transport protocol and the network adapter
card driver so that an NDIS-compliant protocol
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 455
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
455
Select the Appropriate Media
for Various Situations
stack can operate with an NDIS-compliant
adapter driver.
á To define a method for binding multiple proto-
Media choices include:
cols to a single driver so that the adapter can
simultaneously support communications under
multiple protocols. In addition, the method
enables you to bind one protocol to more than
one adapter.
á Twisted-pair cable
á Coaxial cable
á Fiber-optic cable
The Open Data-Link Interface (ODI), developed by
Apple and Novell, serves the same function as NDIS.
Originally, ODI was written for NetWare and
Macintosh environments. Like NDIS, ODI provides
rules that establish a vendor-neutral interface between
the protocol stack and the adapter driver. This interface
also enables one or more network drivers to support
one or more protocol stacks.
á Wireless
Situational elements include:
á Cost
á Distance limitations
á Number of nodes
Summary Table 1 outlines the characteristics of the
cable types discussed in this section.
PLANNING
The planning section on the exam tests your ability to
apply networking components and standards when
designing a network.
Summary Table 2 compares the different types of wireless communication media in terms of cost, ease of
installation, distance, and other issues.
SUMMARY TABLE 1
C O M PA R I S O N
OF
CABLE MEDIA
Cable Type
Cost
Installation
Capacity
Range
EMI
Coaxial Thinnet
Less than STP
Inexpensive/easy
10 Mbps typical
185 m
Less sensitive than UTP
Coaxial Thicknet
Greater than STP
Less than Fiber
Easy
10 Mbps typical
500 m
Less sensitive than UTP
Shielded TwistedPair (STP)
Greater than UTP
Less than Thicknet
Fairly easy
16 Mbps typical
up to 500 Mbps
100 m
typical
Less sensitive than UTP
Unshielded twistedpair (UTP)
Lowest
Inexpensive/easy
10 Mbps typical
up to 100 Mbps
100 m
typical
Most sensitive
Fiber-optic
Highest
Expensive/
Difficult
100 Mbps typical
Tens of
Kilometers
Insensitive
15 918-1 FastFacts.i Network
456
9/1/98 6:58 AM
Page 456
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
SUMMARY TABLE 2
C O M PA R I S O N
OF
WIRELESS MEDIA
Cable Type
Cost
Installation
Distance
Other Issues
Infrared
Cheapest of all the
wireless
Fairly easy; may
require line of sight
Under a kilometer
Can attenuate due to fog and rain
Laser
Similar to infrared
Requires line of site
Can span several
kilometers
Can attenuate due to fog and rain
Narrow band radio
More expensive than
infrared and laser;
may need FCC
license
Requires trained
technicians and can
involve tall radio
towers
Can span hundreds
of kilometers
Low power devices can attenuate;
can be eavesdropped upon; can also
attenuate due to fog, rain, and solar flares
Spread spectrum
radio
More advanced
technology than
narrow band radio,
thus more expensive
Requires trained
technicians and can
involve tall radio
towers
Can span hundreds
of kilometers
Low power devices can attenuate;
can also attenuate due to fog, rain, and
solar flares
Microwave
Very expensive as it
requires link to
satellites often
Requires trained
technicians and can
involve satellite dishes
Can span thousands
of kilometers
Can be eavesdropped upon; can also
attenuate due to fog, rain, and solar flares
Select the Appropriate
Topology for Various TokenRing and Ethernet Networks
The following four topologies are implemented by
Ethernet and token-ring networks:
á Ring. Ring topologies are wired in a circle. Each
node is connected to its neighbors on either side,
and data passes around the ring in one direction
only. Each device incorporates a receiver and a
transmitter and serves as a repeater that passes the
signal to the next device in the ring. Because the
signal is regenerated at each device, signal degeneration is low. Most ring topologies are logical,
and implemented as physical stars. Token-ring
networks follow a ring topology.
á Bus. Star topologies require that all devices con-
nect to a central hub. The hub receives signals
from other network devices and routes the signals
to the proper destinations. Star hubs can be interconnected to form tree or hierarchical network
topologies. A star physical topology is often used
to physically implement a bus or ring logical
topology that is used by both Ethernet and
token-ring networks.
á Star. Star topologies require that all devices con-
nect to a central hub. The hub receives signals
from other network devices and routes the signals
to the proper destinations. Star hubs can be interconnected to form tree or hierarchical network
topologies. A star physical topology is often used
to physically implement a bus or ring logical
topology that is used by both Ethernet and
token-ring networks.
á Mesh. A mesh topology is really a hybrid model
representing a physical topology because a mesh
topology can incorporate all of the previous
topologies. The difference is that in a mesh
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 457
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
topology every device is connected to every other
device on the network. When a new device is
added, a connection to all existing devices must
be made. Mesh topologies can be used by both
Ethernet and token-ring networks.
Select the Appropriate
Network and Transport
Protocol or Protocols for
Various Token-Ring and
Ethernet Networks
Protocol choices include:
á DLC
á AppleTalk
á IPX
á TCP/IP
á NFS
á SMB
Data Link Control (DLC)
The Data Link Control (DLC) protocol does not provide a fully functioning protocol stack. In Windows
NT systems, DLC is used primarily to access to
Hewlett-Packard JetDirect network-interface printers.
DLC also provides some connectivity with IBM mainframes. It is not a protocol that can be used to connect
Windows NT or 95 computers together.
AppleTalk
AppleTalk is the computing architecture developed by
Apple Computer for the Macintosh family of personal
457
computers. Although AppleTalk originally supported
only Apple’s proprietary LocalTalk cabling system, the
suite has been expanded to incorporate both Ethernet
and token-ring Physical layers. Within Microsoft operating systems, AppleTalk is only supported by
Windows NT Server. Windows NT Workstation and
Windows 95 do not support AppleTalk. AppleTalk cannot be used for Microsoft to Microsoft operating system communication, only by NT servers supporting
Apple clients.
The LocalTalk, EtherTalk, and TokenTalk Link Access
Protocols (LLAP, ELAP, and TLAP) integrate
AppleTalk upper-layer protocols with the LocalTalk,
Ethernet, and token-ring environments.
Apple’s Datagram Deliver Protocol (DDP) is a Network
layer protocol that provides connectionless service
between two sockets. The AppleTalk Transaction
Protocol (ATP) is a connectionless Transport layer protocol. Reliable service is provided through a system of
acknowledgments and retransmissions. The AppleTalk
File Protocol (AFP) provides file services and is responsible for translating local file service requests into formats
required for network file services. AFP directly translates command syntax and enables applications to perform file format translations. AFP is responsible for file
system security and verifies and encrypts logon names
and passwords during connection setup.
IPX
The Internetwork Packet Exchange Protocol (IPX) is a
Network layer protocol that provides connectionless
(datagram) service. (IPX was developed from the XNS
protocol originated by Xerox.) As a Network layer protocol, IPX is responsible for internetwork routing and
maintaining network logical addresses. Routing uses the
RIP protocol (described later in this section) to make
route selections. IPX provides similar functionality as
UDP does in the TCP/IP protocol suite.
15 918-1 FastFacts.i Network
458
9/1/98 6:58 AM
Page 458
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
IPX relies on hardware physical addresses found at
lower layers to provide network device addressing. IPX
also uses sockets, or upper-layer service addresses, to
deliver packets to their ultimate destinations. On the
client, IPX support is provided as a component of the
older DOS shell and the current DOS NetWare
requester.
TCP/IP
TCP/IP is a broad protocol that covers many different
areas. This summary presents some of the most important protocols within the TCP/IP protocol suite.
Internet Protocol (IP)
The Internet Protocol (IP) is a connectionless protocol
that provides datagram service, and IP packets are most
commonly referred to as IP datagrams. IP is a packetswitching protocol that performs the addressing and
route selection.
IP performs packet disassembly and reassembly as
required by packet size limitations defined for the Data
Link and Physical layers being implemented. IP also
performs error checking on the header data using a
checksum, although data from upper layers is not errorchecked.
Transmission Control Protocol (TCP)
The Transmission Control Protocol (TCP) is an internetwork connection-oriented protocol that corresponds to
the OSI Transport layer. TCP provides full-duplex,
end-to-end connections. When the overhead of end-toend communication acknowledgment isn’t required, the
User Datagram Protocol (UDP) can be substituted for
TCP at the Transport (host-to-host) level. TCP and
UDP operate at the same layer.
TCP corresponds to SPX in the NetWare environment
(see the NetWare IPX/SPX section). TCP maintains a
logical connection between the sending and receiving
computer systems. In this way, the integrity of the
transmission is maintained. TCP detects any problems
in the transmission quickly and takes action to correct
them. The tradeoff is that TCP isn’t as fast as UDP, due
to the number of acknowledgments received by the
sending host.
TCP also provides message fragmentation and reassembly and can accept messages of any length from upperlayer protocols. TCP fragments message streams into
segments that can be handled by IP. When used with
IP, TCP adds connection-oriented service and performs
segment synchronization, adding sequence numbers at
the byte level.
Windows Internet Naming Services
(WINS)
Windows Internet Naming Service (WINS) provides a
function similar to that of DNS, with the exception
that it provides a NetBIOS name to IP address resolution. This is important because all of Microsoft’s networking requires the capability to reference NetBIOS
names. Normally NetBIOS names are obtained with
the issuance of broadcasts, but because routers normally
do not forward broadcasts, a WINS server is one alternative that can be used to issue IP addresses to
NetBIOS name requests. WINS servers replace the
need for LMHOSTS files on a computer.
Domain Name System (DNS)
The Domain Name System (DNS) protocol provides
host name and IP address resolution as a service to
client applications. DNS servers enable humans to use
logical node names, utilizing a fully qualified domain
name structure to access network resources. Host
names can be up to 260 characters long. DNS servers
replace the need for HOSTS files on a computer.
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 459
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
Network File System (NFS)
Network File System (NFS), developed by Sun
Microsystems, is a family of file-access protocols that
are a considerable advancement over FTP and Telnet.
Since Sun made the NFS specifications available for
public use, NFS has achieved a high level of popularity.
Server Messaging Blocks (SMB)
One protocol that is slightly independent is Microsoft’s
Server Messaging Blocks (SMB). SMBs are Microsoft’s
equivalent to NCP packets. Like NCP packets, SMBs
operate at the Application layer of the OSI model.
SMBs allow machines on a Microsoft network to communicate with one another. Through the use of SMBs,
file and print services can be shared. SMBs can use
TCP/IP, NWLink (IPX/SPX), and NetBEUI because
SMBs utilize a NetBIOS interface when communicating. For more information on NetBIOS names, see the
following section.
459
routers can determine the best path for a packet
based on routing algorithms.
á Brouters. A brouter is a device that is a combina-
tion of a bridge and a router, providing both
types of services.
á Gateways. Gateways function under a process
similar to routers except that gateways can connect dissimilar network environments. A gateway
replaces the necessary protocol layers of a packet
so that the packet can circulate in the destination
environment.
List the Characteristics,
Requirements, and Appropriate
Situations for WAN Connection
Services
WAN connection services include:
Select the Appropriate
Connectivity Devices for
Various Token-Ring and
Ethernet Networks
Connectivity devices include:
á Repeaters. Repeaters regenerate a signal and are
used to expand LANs beyond cabling limits.
á Bridges. Bridges know the side of the bridge on
which a node is located. A bridge passes only
packets addressed to computers across the bridge,
so a bridge can thus filter traffic, reducing the
load on the transmission medium.
á Routers. Routers forward packets based on a log-
ical (as opposed to a physical) address. Some
á X.25
á ISDN
á Frame relay
á ATM
X.25
X.25 is a packet-switching network standard developed
by the International Telegraph and Telephone
Consultative Committee (CCITT), which has been
renamed the International Telecommunications Union
(ITU). The standard, referred to as Recommendation
X.25, was introduced in 1974 and is now implemented
most commonly in WANs.
At the time X.25 was developed, this flow control
and error checking was essential because X.25 was
15 918-1 FastFacts.i Network
460
9/1/98 6:58 AM
Page 460
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
developed around relatively unreliable telephone line
communications. The drawback is that error checking
and flow control slow down X.25. Generally, X.25 networks are implemented with line speeds up to 64 Kbps,
although actual throughput seems slower due to the
error correction controls in place. These speeds are suitable for the file transfer and terminal activity that comprised the bulk of network traffic when X.25 was
defined, most of this traffic being terminal connections
to mainframes. Such speeds, however, are inadequate to
provide LAN-speed services, which typically require
speeds of 1 Mbps or better. X.25 networks, therefore,
are poor choices for providing LAN application services
in a WAN environment. One advantage of X.25, however, is that it is an established standard that is used
internationally. This, as well as lack of other services
throughout the world, means that X.25 is more of a
connection service to Africa, South America, and Asia,
where a lack of other services prevails.
ISDN
The original idea behind ISDN was to enable existing
phone lines to carry digital communications, and it was
at one time touted as a replacement to traditional analog lines. Thus, ISDN is more like traditional telephone service than some of the other WAN services.
ISDN is intended as a dial-up service and not as a permanent 24-hour connection.
ISDN separates the bandwidth into channels. Based
upon how these channels are used, ISDN can be separated into two classes of service:
á Basic Rate (BRI). Basic Rate ISDN uses three
channels. Two channels (called B channels) carry
the digital data at 64 Kbps. A third channel
(called the D channel) provides link and signaling
information at 16 Kbps. Basic Rate ISDN thus is
referred to as 2B+D. A single PC transmitting
through ISDN can use both B channels simultaneously, providing a maximum data rate of 128
Kbps (or higher with compression).
á Primary Rate (PRI). Primary Rate supports 23
64 Kbps B channels and one 64 Kbps D channel.
The D channel is used for signaling and management, whereas the B channels provide the data
throughput.
In a BRI line, if the line was currently being used for
voice, this would only allow one of the B channels to
be available for data. This effectively reduces the
throughput of the BRI to 64 Kbps.
Frame Relay
Frame Relay was designed to support the Broadband
Integrated Services Digital Network (B-ISDN), which
was discussed in the previous section. The specifications
for Frame Relay address some of the limitations of
X.25. As with X.25, Frame Relay is a packet-switching
network service, but Frame Relay was designed around
newer, faster fiber-optic networks.
Unlike X.25, Frame Relay assumes a more reliable network. This enables Frame Relay to eliminate much of
the X.25 overhead required to provide reliable service
on less reliable networks. Frame Relay relies on higherlevel protocol layers to provide flow and error control.
Frame Relay typically is implemented as a public data
network and, therefore, is regarded as a WAN protocol.
The scope of Frame Relay, with respect to the OSI
model, is limited to the Physical and Data Link layers.
Frame Relay provides permanent virtual circuits that
supply permanent virtual pathways for WAN connections. Frame Relay services typically are implemented at
line speeds from 56 Kbps up to 1.544 Mbps (T1).
Customers typically purchase access to a specific
amount of bandwidth on a frame-relay service. This
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 461
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
bandwidth is called the committed information rate
(CIR), a data rate for which the customer is guaranteed
access. Customers might be permitted to access higher
data rates on a pay-per-use temporary basis. This
arrangement enables customers to tailor their network
access costs based on their bandwidth requirements.
To use Frame Relay, you must have special Frame
Relay-compatible connectivity devices (such as framerelay-compatible routers and bridges).
Asynchronous Transfer Mode (ATM)
Asynchronous Transfer Mode (ATM) is a high-bandwidth
switching technology developed by the ITU
Telecommunications Standards Sector (ITU-TSS). An
organization called the ATM Forum is responsible for
defining ATM implementation characteristics. ATM
can be layered on other Physical layer technologies,
such as Fiber Distributed Data Interface (FDDI) and
SONET.
Several characteristics distinguish ATM from other
switching technologies. ATM is based on fixed-length
53-byte cells, whereas other technologies employ frames
that vary in length to accommodate different amounts
of data. Because ATM cells are uniform in length,
switching mechanisms can operate with a high level of
efficiency. This high efficiency results in high data
transfer rates. Some ATM systems can operate at an
incredible rate of 622 Mbps; a typical working speed
for an ATM is around 155 Mbps.
The unit of transmission for ATM is called a cell. All
cells are 53 bytes long and consist of a 5-byte header
and 48 bytes of data. The 48-byte data size was selected
by the standards committee as a compromise to suit
both audio- and data-transmission needs. Audio information, for instance, must be delivered with little latency (delay) to maintain a smooth flow of sound. Audio
engineers therefore preferred a small cell so that cells
461
would be more readily available when needed. For data,
however, large cells reduce the overhead required to
deliver a byte of information.
Asynchronous delivery is another distinguishing feature
of ATM. “Asynchronous” refers to the characteristic of
ATM in which transmission time slots don’t occur periodically but are granted at irregular intervals. ATM uses
a technique called label multiplexing, which allocates
time slots on demand. Traffic that is time-critical, such
as voice or video, can be given priority over data traffic
that can be delayed slightly with no ill effect. Channels
are identified by cell labels, not by specific time slots. A
high-priority transmission need not be held until its
next time slot allocation. Instead, it might be required
to wait only until the current 53-byte cell has been
transmitted.
IMPLEMENTATION
The Implementation section of the exam tests your
knowledge of how to implement, test, and manage an
installed network.
Choosing an Administrative
Plan to Meet Specified Needs,
Including Performance
Management, Account
Management, and Security
Administrative plans can be broken down into three
areas: performance management, account management,
and security.
15 918-1 FastFacts.i Network
462
9/1/98 6:58 AM
Page 462
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
Performance Management
Performance management is best done through the
establishment of a baseline of the network performance
and a baseline of a computer’s performance. Based
upon the information in a baseline, the administrators
of the network can establish when network or computer performance is abnormal.
Choosing a Disaster Recovery
Plan for Various Situations
Disaster recovery applies to many different components
on the network. The following sections describe the
most common issues and solutions used in a disaster
recovery program.
Account Management
Uninterruptible Power Supply (UPS)
Account management within Windows NT is done
through the use of groups. In a workgroup model,
there exist local groups, or groups that are local to the
computer. These groups are not seen on other machines
in the network. Users are placed into these local groups
and assigned permissions to resources, such as printers,
shares, or files and directories.
An uninterruptible power supply (UPS) is a special battery (or sometimes a generator) that supplies power to
an electronic device in the event of a power failure.
UPSs commonly are used with network servers to prevent a disorderly shutdown by warning users to log out.
After a predetermined waiting period, the UPS software
performs an orderly shutdown of the server. Many UPS
units also regulate power distribution and serve as protection against power surges. Remember that in most
cases, a UPS generally does not provide for continued
network functionality for longer than a few minutes. A
UPS is not intended to keep the server running
through a long power outage, but rather to give the
server time to do what it needs before shutting down.
This can prevent the data loss and system corruption
that sometimes result from sudden shutdown.
Windows 95 computers do not have built-in groups.
There also is no account database on a Windows 95
computer to provide user accounts.
Windows NT domain models do make use of user
accounts and groups. Like the workgroup model, the
domain model has user accounts and local groups. A
domain model also has global groups. Global groups
reside on a domain controller and can be referenced as
a resource user by any Windows NT computer within
the domain sharing resources.
Security
Windows 95 computers have the capability to provide
share-level security, which involves password protecting
resources.
Windows NT computers can provide user-level security, in which users are granted access to resources on a
user or local group basis (workgroups and domains support this) and a global group basis (only domains support this).
Tape Backup
Tape backups are done to store data offline in the event
that the hard drive containing the data fails. There are
three types of tape backups:
á Full backup. Backs up all specified files.
á Incremental backup. Backs up only those files
that have changed since the last backup.
á Differential backup. Backs up the specified files
if the files have changed since the last backup.
This type doesn’t mark the files as having been
backed up, however. (A differential backup is
15 918-1 FastFacts.i Network
9/1/98 6:58 AM
Page 463
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
somewhat like a copy command. Because the file
is not marked as having been backed up, a later
differential or incremental backup will back up
the file again.)
463
disks fail, the process can be reversed and any disk can
be reconstructed from the data on the other two.
Recovery includes replacing the bad disk and then
regenerating its data through the Disk Administrator. A
maximum of 32 disks can be connected in a RAID 5
array under Windows NT.
RAID 1
In level 1, drives are paired or mirrored with each byte
of information being written to each identical drive.
You can duplex these devices by adding a separate drive
controller for each drive. Disk mirroring is defined as
two hard drives (one primary, one secondary) that use
the same disk channel or controller cards and cable.
Disk mirroring is most commonly configured by using
disk drives contained in the server. Duplexing is a form
of mirroring that involves the use of a second controller
and that enables you to configure a more robust hardware environment.
Given the Manufacturer’s
Documentation for the Network
Adapter, Install, Configure, and
Resolve Hardware Conflicts for
Multiple Network Adapters in a
Token-Ring or Ethernet
Network
RAID 5
The following resources are configurable on network
adapter cards:
RAID 5 uses striping with parity information written
across multiple drives to enable fault-tolerance with a
minimum of wasted disk space. This level also offers
the advantage of enabling relatively efficient performance on writes to the drives, as well as excellent read
performance.
Striping with parity is based on the principle that all
data is written to the hard drive in binary code (ones
and zeros). RAID 5 requires at least three drives
because this version writes data across two of them and
then creates the parity block on the third. If the first
byte is 00111000 and the second is 10101001, the system computes the third by adding the digits together
using this system:
1+1=0, 0+0=0, 0+1=1, 1+0=1
The sum of 00111000 and 10101001 is 10010001,
which would be written to the third disk. If any of the
á IRQ
á Base I/O port address
á Base memory address
á DMA channel
á Boot PROM
á MAC address
á Ring speed (token-ring cards)
á Connector type
Not all network adapter cards have all of these
resources available for configuration. These resource
settings on the network adapter card must be different
than the settings found on other components used
within the computer.
Some network adapter cards use jumper settings to
configure these settings, others use software, and others
15 918-1 FastFacts.i Network
464
9/1/98 6:58 AM
Page 464
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
can have this done through the operating system software, such as Windows 95 and Windows NT. The
method of configuration is dependent upon the manufacturer.
Implementing a NetBIOS
Naming Scheme for All
Computers on a Given Network
NetBIOS is an interface that provides NetBIOS-based
applications with access to network resources. Every
computer on a Windows NT network must have a
unique name for it to be accessible through the
NetBIOS interface. This unique name is called a computer name or a NetBIOS name.
On a NetBIOS network, every computer must have a
unique name. The computer name can be up to 15
characters long. A NetBIOS name can include alphanumeric characters and any of the following special characters:
[email protected]#$%^&()-_’{}.~
Note that you cannot use a space or an asterisk in a
NetBIOS name. Also, NetBIOS names are not case
sensitive.
Selecting the Appropriate
Hardware and Software Tools
to Monitor Trends in the
Network
The hardware and software tools described in the next
five sections are used to monitor trends in a network.
Protocol Analyzer
This can be a hardware or software tool to analyze the
traffic in a network. Protocol analyzers capture packets
on a network and display their contents. The software
version of this tool supplied by Microsoft is Network
Monitor. Network Monitor ships with Windows NT as
a scaled-down version that can only capture data
between the host computer and those to which the host
talks.
Event Viewer
This software tool is found on Windows NT. It reports
one of three event types:
á System Events. Those generated by the operating
system.
á Application Events. Those generated by any
application that is programmed to make event
calls to the Event Viewer.
á Auditing. Any auditing being performed on
NTFS partitions or by users interacting with the
network.
Performance Monitor
Windows NT’s Performance Monitor tool lets you
monitor important system parameters for the computers on your network in real time. Performance Monitor
can keep an eye on a large number of system parameters, providing a graphical or tabular profile of system
and network trends. Performance Monitor also can save
performance data in a log for later reference. You can
use Performance Monitor to track statistical measurements (called counters) for any of several hardware or
software components (called objects).
15 918-1 FastFacts.i Network
9/1/98 6:59 AM
Page 465
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
465
System Monitor
Protocol Analyzers
Windows 95 includes a program called System Monitor
that also allows information to be collected on the
Windows 95 machine in real time. System Monitor
collects information on different categories of items on
the system. System Monitor is not as detailed as
Windows NT’s Performance Monitor.
Protocol analyzers are either hardware or software products used to monitor network traffic, track network
performance, and analyze packets. Protocol analyzers
can identify bottlenecks, protocol problems, and malfunctioning network components.
Digital Volt Meter (DVM)
Simple Network Management
Protocol (SNMP)
SNMP is a TCP/IP protocol used to perform management operations on a TCP/IP network. SNMP-enabled
devices allow for information to be sent to a management utility (this is called a trap). SNMP devices also
allow for the setting and extraction of information (this
is done by the issuance of a set or get command)
found in their Management Information Base (MIB).
TROUBLESHOOTING
The Troubleshooting section of the exam covers many
of the topics covered in previous sections. Emphasis of
this section is to test your understanding of what can
cause problems, and how to fix them.
Identifying Common Errors
Associated with Components
Required for Communications
The utilities described in the next four sections can be
used to diagnose errors associated with components
required for communications.
Digital volt meters are handheld electronic measuring
tools that enable you to check the voltage of network
cables. They also can be used to check the resistance of
terminators. You can use a DVM to help you find a
break or a short in a network cable.
DVMs are usually inexpensive battery-operated devices
that have either a digital or needle readout and two
metal prongs attached to the DVM by some wires a
foot or more in length. By sending a small current
through the wires and out through the metal prongs,
resistance and voltages of terminators and wires can be
measured.
Time-Domain Reflectometers (TDR)
Time-domain reflectometers send sound waves along a
cable and look for imperfections that might be caused
by a break or a short in the line. A good TDR can
detect faults on a cable to within a few feet.
Oscilloscope
An oscilloscope measures fluctuations in signal voltage
and can help find faulty or damaged cabling.
Oscilloscopes are often more expensive electronic
devices that show the signal fluctuations on a monitor.
Several diagnostic software tools provide information
on virtually any type of network hardware, as well. A
considerable number of diagnostic software packages
are available for a variety of prices.
15 918-1 FastFacts.i Network
466
9/1/98 6:59 AM
Page 466
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
A common software tool distributed with most network cards is a Send/Receive package. This software
tool allows two computers with network cards and
cables to connect to each other. This tool does not rely
on a networked operating system, nor can it be used to
send data. It simply sends packets from one computer
to the other, establishing that the network cards and
underlying transmission media are connected and configured properly.
Diagnosing and Resolving
Common Connectivity
Problems with Cards, Cables,
and Related Hardware
Most network problems occur on the transmission
media or with the components that attach devices to
the transmission media. All of these components operate at the Physical, DataLink, or Network levels of the
OSI model. The components that connect PCs and
enable them to communicate are susceptible to many
kinds of problems.
Troubleshooting Cables and
Connectors
Most network problems occur at the OSI Physical
layer, and cabling is one of the most common causes. A
cable might have a short or a break, or it might be
attached to a faulty connector. Tools such as DVMs
and TDRs help search out cabling problems.
Cabling problems can cause three major problems: An
individual computer cannot access the network, a
group of computers cannot access the network, or none
of the computers can access the network.
On networks that are configured in a star topology, an
individual cable break between the computer and hub
or MSAU causes a failure in communication between
that individual computer and the rest of the network.
This type of cable break does not cause problems
between all of the other computers on the network.
A cable break in cables connecting multiple hubs causes
a break in communications between the computers on
one side of the cable break and the computers on the
other side of the cable break. In most cases, the communications between computers within the broken segment can continue.
In the case of MSAU, the breakage of a cable connecting MSAUs often causes all computers on the ring to
fail because the ring is not complete. A break in the
cable on a bus topology also causes all computers on
the network segment to be unable to communicate
with any other computers on the network.
Try the following checks when troubleshooting network cabling problems:
á With 10BASE-T, make sure the cable used has
the correct number of twists to meet the datagrade specifications.
á Look for electrical interference, which can be
caused by tying the network cable together with
monitor and power cords. Fluorescent lights,
electric motors, and other electrical devices can
cause interference if they are located too close to
cables. These problems often can be alleviated by
placing the cable away from devices that generate
electromagnetic interference or by upgrading the
cable to one that has better shielding.
á Make sure that connectors are pinned properly
and crimped tightly.
á If excess shielding on coaxial cable is exposed,
make sure it doesn’t ground out the connector.
á Ensure that coaxial cables are not coiled tightly
together. This can generate a magnetic field
around the cable, causing electromagnetic interference.
15 918-1 FastFacts.i Network
9/1/98 6:59 AM
Page 467
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
á On coaxial Ethernet LANs, look for missing ter-
minators or terminators with improper resistance
ratings.
á Watch out for malfunctioning transceivers, con-
centrators, or T-connectors. All of these components can be checked by replacing the suspect
devices.
á Test the continuity of the cable by using the vari-
ous physical testing devices discussed in the previous section or by using a software-based cable
testing utility.
á Make sure that all the component cables in a seg-
ment are connected. A user who moves his client
and removes the T-connector incorrectly can
cause a broken segment.
á Examine cable connectors for bent or broken
pins.
á On token-ring networks, inspect the attachment
of patch cables and adapter cables. Remember,
patch cables connect MSAUs, and adapter cables
connect the network adapter to the MSAU.
One advantage of a token-ring network is its built-in
capability to monitor itself. token-ring networks provide electronic troubleshooting and, when possible,
actually make repairs. When the token-ring network
can’t make its own repairs, a process called beaconing
narrows down the portion of the ring in which the
problem is most likely to exist.
467
á Make sure the cable is properly connected to the
card.
á Confirm that you have the correct network
adapter card driver and that the driver is installed
properly. Be sure the card is properly bound to
the appropriate transport protocol.
á Make sure the network adapter card and the net-
work adapter card driver are compatible with
your operating system. If you use Windows NT,
consult the Windows NT hardware compatibility
list. If you use Windows 95 or another operating
system, rely on the adapter card vendor specifications.
á Test for resource conflicts. Make sure another
device isn’t attempting to use the same resources.
If you think a resource conflict might be the
problem, but you can’t pinpoint the conflict
using Windows NT Diagnostics, Windows 95’s
Device Manager, or some other diagnostic program, try removing all the cards except the network adapter and then replacing the cards one by
one. Check the network with each addition to
determine which device is causing the conflict.
á Run the network adapter card’s diagnostic soft-
ware. This will often indicate which resource on
the network card is failing.
á Examine the jumper and DIP switch settings on
the card. Make sure the resource settings are consistent with the settings configured through the
operating system.
Troubleshooting Network Adapter
Cards
á Make sure the card is inserted properly in the
Network problems often result from malfunctioning
network adapter cards. The process of troubleshooting
the network adapter works like any other kind of troubleshooting process: Start with the simple. The following list details some aspects you can check if you think
your network adapter card might be malfunctioning:
á If necessary, remove the card and clean the con-
slot. Reseat if necessary.
nector fingers (don’t use an eraser because it
leaves grit on the card).
á Replace the card with one that you know works.
If the connection works with a different card, you
know the card is the problem.
15 918-1 FastFacts.i Network
468
9/1/98 6:59 AM
Page 468
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
Token-ring network adapters with failure rates that
exceed a preset tolerance level might actually remove
themselves from the network. Try replacing the card.
Some token-ring networks also can experience problems if a token-ring card set at a ring speed of 16 Mbps
is inserted into a ring using a 4 Mbps ring speed, and
vice versa.
MSAU port is a quick fix that is useful on relatively
small token-ring networks.
Isolating problems with patch cables, adapter cables,
and MSAUs is easier to do if you have a current log of
your network’s physical design. After you narrow down
the problem, you can isolate potential problem areas
from the rest of the network and then use a cable tester
to find the actual problem.
Troubleshooting Hubs and MSAUs
If you experience problems with a hub-based LAN,
such as a 10BASE-T network, you often can isolate the
problem by disconnecting the attached workstations
one at a time. If removing one of the workstations
eliminates the problem, the trouble may be caused by
that workstation or its associated cable length. If
removing each of the workstations doesn’t solve the
problem, the fault may lie with the hub. Check the
easy components first, such as ports, switches, and connectors, and then use a different hub (if you have it) to
see if the problem persists. If your hub doesn’t work
properly, call the manufacturer.
If you’re troubleshooting a token-ring network, make
sure the cables are connected properly to the MSAUs,
with ring-out ports connecting to the ring-in ports
throughout the ring. If you suspect the MSAU, isolate
it by changing the ring-in and ring-out cables to bypass
the MSAU. If the ring is now functional again, consider replacing the MSAU. In addition, you might find
that if your network has MSAUs from more than one
manufacturer, they are not wholly compatible.
Impedance and other electrical characteristics can show
slight differences between manufacturers, causing intermittent network problems. Some MSAUs (other than
the 8228) are active and require a power supply. These
MSAUs fail if they have a blown fuse or a bad power
source. Your problem also might result from a misconfigured MSAU port. MSAU ports using the hermaphrodite connector need to be reinitialized with the setup
tool. Removing drop cables and reinitializing each
Troubleshooting Modems
A modem presents all the potential problems you find
with any other device. You must make sure that the
modem is properly installed, that the driver is properly
installed, and that the resource settings do not conflict
with other devices. Modems also pose some unique
problems because they must connect directly to the
phone system, they operate using analog communications, and they must make a point-to-point connection
with a remote machine.
The online help files for both Windows NT and
Windows 95 include a topic called the Modem
Troubleshooter. The Modem Troubleshooter leads you
to possible solutions for a modem problem by asking
questions about the symptoms. As you answer the
questions (by clicking the gray box beside your answer),
the Modem Troubleshooter zeroes in on more specific
questions until (ideally) it leads you to a solution.
Some common modem problems are as follows:
á Dialing problems. The dialing feature is improp-
erly configured. For instance, the modem isn’t
dialing 9 to bypass your office switchboard, or it
is dialing 9 when you’re away from your office.
The computer also could be dialing an area code
or an international code when it shouldn’t. Check
the dialing properties for the connection.
á Connection problems. You cannot connect to
another modem. Your modem and the other
modem might be operating at different speeds.
15 918-1 FastFacts.i Network
9/1/98 6:59 AM
Page 469
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
Verify that the maximum speed setting for your
modem is the highest speed that both your
modem and the other modem can use. Also make
sure the Data Bits, Parity, and Stop Bits settings
are consistent with the remote computer.
á Digital phone systems. You cannot plug a
modem into a telephone line designed for use
with digital phone systems. These digital phone
systems are commonplace in most office environments.
á Protocol problems. The communicating devices
are using incompatible line protocols. Verify that
the devices are configured for the same or compatible protocols. If one computer initiates a connection using PPP, the other computer must be
capable of using PPP.
Repeaters, Bridges, and Routers
Issues dealing with repeaters, bridges, and routers are
often more technically advanced than those covered in
a book such as Networking Essentials. Companies such
as Cisco, Bay Networks, and 3Com have their own
dedicated books and courses on dealing with the installation, configuration, and troubleshooting of repeaters,
bridges, and routers. In general, there are some basic
troubleshooting steps you can do when working with
these three devices.
Repeaters are responsible for regenerating a signal sent
down the transmission media. The typical problem
with repeaters is that they do not work—that is, the
signal is not being regenerated. If this is the case, the
signal being sent to devices on the other side of the
repeater from the sending device will not receive the
signal.
Problems with bridges are almost identical to that of a
repeater. The signal being sent to devices on the other
side of the bridge from the sending device will be
received. Other issues with bridges are that the table of
469
which devices are on which interface of the bridge can
get corrupt. This can lead from one to all machines not
receiving packets on the network. Diagnostic utilities
provided by the bridge’s manufacturer can resolve this
type of problem.
Problems with routers can be complex, and troubleshooting them often involves a high level of understanding of the different protocols in use on the
network, as well as the software and commands used
to program a router. There are generally two types of
router problems.
The first router problem that is commonly found is
that packets are just not being passed through because
the router is ‘dead’ or simply not functioning. The second common problem with routers is that the routing
tables within the routers are corrupted or incorrectly
programmed. This problem either leads to computers
on different networks being unable to communicate
with each other or to the fact that certain protocols
simply do not work.
Resolve Broadcast Storms
A broadcast storm is a sudden flood of broadcast messages that clogs the transmission medium, approaching
100 percent of the bandwidth. Broadcast storms cause
performance to decline and, in the worst case, computers cannot even access the network. The cause of a
broadcast storm is often a malfunctioning network
adapter, but a broadcast storm also can be caused when
a device on the network attempts to contact another
device that either doesn’t exist or for some reason
doesn’t respond to the broadcast.
If the broadcast messages are viable, a networkmonitoring or protocol-analysis tool often can determine the source of the storm. If the broadcast storm is
caused by a malfunctioning adapter throwing illegible
packets onto the line, and a protocol analyzer can’t find
the source, try to isolate the offending PC by removing
15 918-1 FastFacts.i Network
470
9/1/98 6:59 AM
Page 470
FA S T FA C T S : N E T W O R K I N G E S S E N T I A L S E X A M
computers from the network one at a time until the
line returns to normal.
Identify and Resolve Network
Performance Problems
If your network runs slower than it used to run (or
slower than it ought to run), the problem might be that
the present network traffic exceeds the level at which
the network can operate efficiently. Some possible causes for increased traffic are new hardware (such as a new
workstation) or new software (such as a network computer game or some other network application). A generator or another mechanical device operating near the
network could cause a degradation of network performance. In addition, a malfunctioning network device
could act as a bottleneck. Ask yourself what has
changed since the last time the network operated efficiently, and begin there with your troubleshooting
efforts.
A performance monitoring tool, such as Windows NT’s
Performance Monitor or Network Monitor, can help
you look for bottlenecks that are adversely affecting
your network. For instance, the increased traffic could
be the result of increased usage. If usage exceeds the
capacity of the network, you might want to consider
expanding or redesigning your network. You also might
want to divide the network into smaller segments by
using a router or a bridge to reduce network traffic. A
protocol analyzer can help you measure and monitor
the traffic at various points on your network.
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 471
Now that you have thoroughly read through this book,
worked through the exercises and got as much hands
on exposure to NT Server as you could, you’ve now
booked your exam. This chapter is designed as a last
minute cram for you as you walk out the door on your
way to the exam. You can’t re-read the whole book in
an hour, but you will be able to read this chapter in
that time.
This chapter is organized by objective category, giving
you not just a summary, but a rehash of the most
important point form facts that you need to know.
Remember that this is meant to be a review of concepts and a trigger for you to remember wider definitions. In addition to what is in this chapter, make sure
you know what is in the glossary because this chapter
does not define terms. If you know what is in here
and the concepts that stand behind it, chances are the
exam will be a snap.
PLANNING
Remember: Here are the elements that Microsoft says
they test on for the “Planning” section of the exam.
á Plan the disk drive configuration for various
requirements. Requirements include: choosing a
file system and fault tolerance method
á Choose a protocol for various situations.
Protocols include: TCP/IP, NWLink IPX/SPX
Compatible Transport, and NetBEUI
Minimum requirement for installing NT Server on an
Intel machine is 468DX/33, 16MB of RAM, and
130MB of free disk space.
The login process on an NT Domain is as follows:
1. WinLogon sends the user name and password to
the Local Security Authority (LSA).
Fast Facts
WINDOWS NT SERVER 4
EXAM
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 472
472
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
2. The LSA passes the request to the local NetLogon
service.
3. The local NetLogon service sends the logon
information to the NetLogon service on the
domain controller.
4. The NetLogon service on the domain controller
passes the information to the domain controller’s
Security Accounts Manager (SAM).
5. The SAM asks the domain directory database for
approval of the user name and password.
6. The SAM passes the result of the approval
request to the domain controller’s NetLogon
service.
7. The domain controller’s NetLogon service passes
the result of the approval request to the client’s
NetLogon service.
Stripe sets with Parity use between 3 and 32 hard drives
and provides a (n-1)/n*100% utilization (n = number
of disks in the set).
Disk duplexing provides better tolerance than mirroring because it does mirroring with separate controllers
on each disk.
NT Supports 3 file systems: NTFS, FAT, and CDFS
(it no longer supports HPFS, the OS/2 file system nor
does it support FAT32, a file system used by Windows
95).
The following table is a comparison of NTFS and FAT
features:
Table 1.1 shows a quick summary of the differences
between file systems:
SUMMARY TABLE 1
8. The client’s NetLogon service passes the result of
the approval request to the LSA.
FAT
Feature
FAT
NTFS
9. If the logon is approved, the LSA creates an
access token and passes it to the WinLogon
process.
File name length
255
255
8.3 file name compatibility
Yes
Yes
File size
4 GB
16 EB
Partition size
4 GB
16 EB
Directory structure
Linked list
B-tree
Local security
No
Yes
Transaction tracking
No
Yes
Hot fixing
No
Yes
Overhead
1 MB
>4 MB
Required on system partition for
RISC-based computers
Yes
No
Accessible from MS-DOS/ Windows 95
Yes
No
Accessible from OS/2
Yes
No
Case-sensitive
No
POSIX
only
Case preserving
Yes
Yes
10. WinLogon completes the logon, thus creating a
new process for the user and attaching the access
token to the new process.
The system partition is where your computer boots and
it must be on an active partition.
The boot partition is where the WINNT folder is
found and it contains the NT program files. It can be
on any partition (not on a volume set, though).
NT supports two forms of software-based fault tolerance: Disk Mirroring (RAID 1) and Stripe Sets with
Paritiy (RAID 5).
Disk Mirroring uses 2 hard drives and provides 50%
disk space utilization.
VERSUS
N T F S C O M PA R I S O N
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 473
FAST FACTS: WINDOWS NT SERVER 4 EX AM
Feature
FAT
NTFS
Compression
No
Yes
Efficiency
200 MB
400 MB
Windows NT formattable
Yes
Yes
Fragmentation level
High
Low
Floppy disk formattable
Yes
No
The following is a table to summarize the protocols
commonly used by NT for network communication:
SUMMARY TABLE 2
PRIMARY PROTOCOL USES
Protocol
Primary Use
TCP/IP
Internet and WAN connectivity
NWLink
Interoperability with NetWare
NetBEUI
Interoperability with old Lan Man networks
The main points regarding TCP/IP are as follows:
á Requires IP Address, and Subnet Mask to function (default Gateway if being routed)
á Can be configured manually or automatically
using DHCP server running on NT
á Common address resolution methods are WINS
and DNS
INSTALLATION AND
CONFIGURATION
Remember: Here are the elements that Microsoft says
they test on for the “Installation and Configuration”
section of the exam.
473
á Install Windows NT Server on Intel-based platforms.
á Install Windows NT Server to perform various
server roles. Server roles include: Primary
domain controller, Backup domain controller,
and Member server.
á Install Windows NT Server by using various
methods. Installation methods include: CDROM, Over-the-network, Network Client
Administrator, and Express versus custom.
á Configure protocols and protocol bindings.
Protocols include: TCP/IP, NWLink IPX/SPX
Compatible Transport, and NetBEUI.
á Configure network adapters. Considerations
include: changing IRQ, IObase, and memory
addresses and configuring multiple adapters.
á Configure Windows NT server core services.
Services include: Directory Replicator, License
Manager, and Other services.
á Configure peripherals and devices. Peripherals
and devices include: communication devices,
SCSI devices, tape devices drivers, UPS devices
and UPS service, mouse drivers, display drivers,
and keyboard drivers.
á Configure hard disks to meet various requirements. Requirements include: allocating disk
space capacity, providing redundancy, improving
security, and formatting.
á Configure printers. Tasks include: adding and
configuring a printer, implementing a printer
pool, and setting print priorities.
á Configure a Windows NT Server computer for
various types of client computers. Client computer types include: Windows NT Workstation,
Microsoft Windows 95, and Microsoft MSDOS-based.
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 474
474
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
The Hardware Compatibility list is used to ensure that
NT supports all computer components.
NT can be installed in 3 different configurations in a
domain: Primary Domain Controller, Backup Domain
Controller, and Member Server.
Two sources can be used for installation files: CDROM or network share (which is the hardware specific
files from the CD copied onto a server and shared).
Three Setup diskettes are required for all installations
when a CD-ROM is not supported by the operating
system present on the computer at installation time (or
if no operating system exists and the computer will not
boot from the CD-ROM.)
WINNT and WINNT32 are used for network installation; WINNT32 for installations when NT is currently
present on the machine you are installing to and
WINNT when it is not.
The following table is a summary of the WINNT and
WINNT32 switches:
SUMMARY TABLE 3
WI NNT
AND
WINN T32 S W I T C H F U N C T I O N S
Switch
Function
/B
Prevents creation of the three setup disks during the
installation process
/S
Indicates the location of the source files for NT installation (e.g., /S:D:\NTFiles)
/U
Indicates the script file to use for an unattended
installation (e.g., /U:C:\Answer.txt)
/UDF
Indicates the location of the uniqueness database file
which defines unique configuration for each NT
machine being installed (e.g., /UDF:D:\Answer.UDF)
/T
Indicates the place to put the temporary installation
files
/OX
Initiates only the creation of the three setup disks
Switch
Function
/F
Indicates not to verify the files copied to the setup
diskettes
/C
Indicates not to check for free space on the setup
diskettes before creating them
To remove NT from a computer you must do the following:
1. Remove all the NTFS partitions from within
Windows NT and reformat them with FAT (this
ensures that these disk areas will be accessible by
non-NT operating systems).
2. Boot to another operating system, such as
Windows 95 or MS-DOS.
3. Delete the Windows NT installation directory
tree (usually WINNT).
4. Delete pagefile.sys.
5. Turn off the hidden, system, and read-only attributes for NTBOOTDD.SYS, BOOT.INI,
NTLDR, and NTDETECT.COM and then
delete them. You might not have all of these on
your computer, but if so, you can find them all in
the root directory of your drive C.
6. Make the hard drive bootable by placing another
operating system on it (or SYS it with DOS or
Windows 95 to allow the operating system with
does exist to boot).
The Client Administrator allows you to do the following:
á Make Network Installation Startup disk: shares
files and creates bootable diskette for initiating
client installation.
á Make Installation Disk Set: copies installation
files to diskette for installing simple clients like
MS-DOS network client 3.0.
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 475
FAST FACTS: WINDOWS NT SERVER 4 EX AM
á Copy Client-Based Network Administration
Tools: creates a folder which can be attached to
from Windows NT Workstation and Windows
95 clients to install tools for administering an NT
Server from a workstation.
MANAGING RESOURCES
Remember: Here are the elements that Microsoft says
they test on for the “Managing Resources” section of
the exam.
á Manage user and group accounts. Considerations include: managing Windows NT groups,
managing Windows NT user rights, administering account policies, and auditing changes to the
user account database.
á Create and manage policies and profiles for various situations. Policies and profiles include: local
user profiles, roaming user profiles, and system
policies.
á Administer remote servers from various types of
client computers. Client computer types include:
Windows 95 and Windows NT Workstation.
á Manage disk resources. Tasks include: copying
and moving files between file systems, creating
and sharing resources, implementing permissions
and security, and establishing file auditing.
Network properties dialog box lets you install and configure the following:
á Computer and Domain names
á Services
á Protocols
á Adapters
á Bindings
475
When configuring NWLink ensure that if more than
one frame type exists on your network that you don’t
use AutoDetect or only the first frame type encountered will be detected from then on.
The following table shows you three TCP/IP command-line diagnostic tools and what they do:
SUMMARY TABLE 4
T C P /IP C O M M A N D L I N E D I A G N O S T I C T O O L S
Tool
Function
IPConfig
Displays the basic TCP/IP configuration of each
adapter card on a computer (with/all displays
detailed configuration information)
Ping
Determines connectivity with another TCP/IP host
by sending a message that is echoed by the recipient
if received
Tracert
Traces each hop on the way to a TCP/IP host and
indicates points of failure if they exist
Network adapter card configuration of IRQ and I/O
port address may or may not be configurable from the
Network Properties dialog box; it depends on the card.
To allow NT computers to participate in a domain, a
computer account must be created for each one.
Windows 95 clients need special profiles and policies
created on a Windows 95 machine and then copied
onto an NT Server to participate in domain profile
and policy configuration.
Windows 95 clients need printer drivers installed on
an NT Server acting as a print controller to print to
an NT controller printer.
Typical services tested for NT Server are listed and
described in the following table:
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 476
476
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
SUMMARY TABLE 5
NT S E R V E R S E R V I C E S
FUNCTIONS
AND
THEIR
Service
Function
DNS
Provides TCP/IP address resolution using a static
table and can be use for non-Microsoft hosts
WINS
Provides TP/IP address resolution using a dynamic
table and can be used for Microsoft hosts
DHCP
Provides automatic configuration of TCP/IP clients
for Microsoft clients
Browser
Provides a list of domain resources to Network
Neighborhood and Server Manager
Replicator
Provides import and export services for automated
file distribution between NT computers (Servers can
be export and import, Workstations can only be
import)
REGEDT32.EXE and REGEDIT are used to view and
modify registry settings in NT.
The five registry subtrees are:
á HKEY_LOCAL_MACHINE. Stores all the computer-specific configuration data.
á HKEY_USERS. Stores all the user-specific configuration data.
á HKEY_CURRENT_USER. Stores all configuration data for the currently logged on user.
á HKEY_CLASSES_ROOT. Stores all OLE and
file association information.
á HKEY_CURRENT_CONFIG. Stores information about the hardware profile specified at startup.
REGEDT32.EXE allows you to see and set security on
the registry and allows you to open the registry in readonly mode, but does not allow you to search by key
value.
NT checking for serial mice at boot may disable a UPS.
To disable that check, place the /noserialmice in the
boot line in the BOOT.INI file.
The SCSI adapters icon in the Control Panel lets you
add and configure SCSI devices as well as CD-ROM
drives.
Many changes made in the disk administrator require
that you choose the menu Partition, Commit Changes
for them to take effect.
Although you can set drive letters manually, the following is how NT assigns letters to partitions and volumes:
1. Beginning from the letter C:, assign consecutive
letters to the first primary partition on each physical disk.
2. Assign consecutive letters to each logical drive,
completing all on one physical disk before moving on to the next.
3. Assign consecutive letters to the additional primary partitions, completing all on one physical disk
before moving on to the next.
Disk Administrator allows for the creation of two kinds
of partitions (primary and extended) and four kinds of
volumes (volume set, stripe set, mirror set, and stripe
set with parity). The following table is a summary of
their characteristics:
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 477
FAST FACTS: WINDOWS NT SERVER 4 EX AM
477
SUMMARY TABLE 6
PA R T I T I O N C H A R A C T E R I S T I C S
Object
Characteristics
Primary partition
Non-divisible disk unit which can be marked active and can be made bootable.
Can have up to four on a physical drive.
NT system partition must be on a primary.
Extended partition
Divisible disk unit which must be divided into logical disks (or have free space used in a volume) in order to function as space storage tool.
Can have only one on a physical drive.
Logical drive within can be the NT boot partition.
Volume Set
Made up of 2-32 portions of free space which do not have to be the same size and which can be spread out over
between 1 and 32 disks of many types (IDE, SCSI, etc.).
Can be added to if formatted NTFS.
Cannot contain NT boot or system partition.
Removing one portion of the set destroys the volume and the data is lost.
Is not fault tolerant.
Stripe Set
Made up of 2-32 portions of free space which have to be the same size and which can be spread out over between 2
and 32 disks of many types (IDE, SCSI, etc.).
Cannot be added to and removing one portion of the set destroys the volume and the data is lost.
Is not fault tolerant.
Mirror Set
Made up of 2 portions of free space which have to be the same size and which must be on 2 physical disks.
Identical data is written to both mirror partitions and they are treated as one disk.
If one disk stops functioning the other will continue to operate.
The NT Boot and System partitions can be held on a mirror set.
Has a 50% disk utilization rate.
Is fault tolerant.
Stripe Set with Parity
Made up of 3-32 portions of free space which have to be the same size and must be spread out over the same
number of physical disks.
Maintains fault tolerance by creating parity information across a stripe.
If one disk fails, the stripe set will continue to function, albeit with a loss of performance.
The NT Boot and System partitions cannot be held on a Stripe Set with Parity.
Is fault tolerant.
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 478
478
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
Disk Administrator can be used to format partitions
and volumes either FAT or NTFS.
If you have any clients who access a shared printer that
are not using NT or are not using the same hardware
platform as your printer server then you must install
those drivers when you share the printer.
By assigning different priorities for printers associated
with the same print device you can create a hierarchy
among users’ print jobs, thus ensuring that the print
jobs of some users print sooner than others.
By adjusting the printer schedule you can ensure that
jobs sent to particular printers are only printed at certain hours of the day.
A printer has permissions assigned to it. The following
is a list of the permissions for printers.
á No Access. Completely restricts access to the
printer.
á Print. Allows a user or group to submit a print
job, and to control the settings and print status
for that job.
á Manage Documents. Allows a user or group to
submit a print job, and to control the settings
and print status for all print jobs.
á Full Control. Allows a user to submit a print
job, and to control the settings and print status
for all documents as well as for the printer itself.
In addition, the user or group may share, stop
sharing, change permissions for, and even delete
the printer.
Printer pools consist of one or more print devices that
can use the same print driver controlled by a single
printer.
MS-DOS users must have print drivers installed locally
on their computers.
The assignment of permissions to resources should use
the following procedure:
1. Create user accounts.
2. Create global groups for the domain and populate the groups with user accounts.
3. Create local groups and assign them rights and
permissions to resources and programs in the
domain.
4. Place global groups into the local groups you
have created, thereby giving the users who are
members of the global groups access to the system and its resources.
The built-in local groups in a Windows NT Domain
are as follows:
á Administrators
á Users
á Guests
á Backup Operators
á Replicator
á Print Operators
á Server Operators
á Account Operators
The built-in global groups in an NT Domain are as
follows:
á Domain Admins
á Domain Users
á Domain Guests
The system groups on an NT server are as follows:
á Everyone
á Creator Owner
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 479
FAST FACTS: WINDOWS NT SERVER 4 EX AM
479
The following table describes the buttons on the User
Properties dialog box and their functions:
á Network
á Interactive
The built-in users on an NT server are as follows:
á Administrator
á Guest
SUMMARY TABLE 7
BUTTONS
ON THE
USER PROPERTIES DIALOG BOX
Button
Function
Groups
Enables you to add and remove group memberships for the account. The easiest way to grant rights to a user account is to
add it to a group that possesses those rights.
Profile
Enables you to add a user profile path, a logon script name, and a home directory path to the user’s environment profile.
You learn more about the Profile button in the following section.
Hours
Enables you to define specific times when the users can access the account. (The default is always.)
Logon To
Enables you to specify up to 8 workstations from which the user can log on. (The default is all workstations.)
Account
Enables you to provide an expiration date for the account. (The default is never.) You also can specify the account as global
(for regular users in this domain) or domain local.
The following table is a summary of the account policy fields:
SUMMARY TABLE 8
ACCOUNT POLICY FIELDS
Button
Function
Maximum Password Age
The maximum number of days a password can be in effect until it must be changed.
Minimum Password Age
The minimum number of days a password must stay in effect before it can be changed.
Minimum Password Length
The minimum number of characters a password must include.
Password Uniqueness
The number of passwords that NT remembers for a user; these passwords cannot be reused until they are
no longer remembered.
Account Lockout
The number of incorrect passwords that can be input by a user before the account becomes locked. Reset
will automatically set the count back to 0 after a specified length of time. In addition the duration of lockout is either a number of minutes or forever (until an administrator unlocks it).
Forcibly disconnect remote
users from server when logon
hours expire
In conjunction with logon hours, this checkbox enables forcible disconnection of a user when authorized
hours come to a close.
Users must log on in order to
change password
Ensures that a user whose password has expired cannot change his or her password but has to have it reset
by an administrator.
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 480
480
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
Account SIDs are unique; therefore, if an account is
deleted, the permissions cannot be restored by recreating an account with the same name.
Local profiles are only available from the machine on
which they were created, whereas roaming profiles can
be accessed from any machine on the network.
A mandatory profile is a roaming profile that users cannot change. They have the extension .MAN.
Hardware profiles can be used with machines that have
more than one hardware configuration (such as laptops).
policy created on an Windows 95 machine and copied
to an NT machine and named Config.Pol.
The Net Use command line can be used to map a drive
letter to a network share; using the /persistent switch
ensures that it is reconnected at next logon.
FAT long file names under NT have 8.3 aliases created
to ensure backward compatibility. The following is an
example of how aliases are generated from 5 files that
all have the same initial characters:
Team meeting Report #3.doc
TEAMME~1.DOC
The System Policy editor (POLEDIT) has two modes,
Policy File mode and Registry Mode.
Team meeting Report #4.doc
TEAMME~2.DOC
Team meeting Report #5.doc
TEAMME~3.DOC
The application of system policies is as follows:
Team meeting Report #6.doc
TEAMME~4.DOC
Team meeting Report #7.doc
TE12B4~1.DOC
1. When you log in, the NT Config.pol is checked.
If there is an entry for the specific user, then any
registry settings indicated will be merged with,
and overwrite if necessary, the users registry.
2. If there is no specific user entry, any settings for
groups that the user is a member of will be
applied to the user.
3. If the user is not present in any groups and not
listed explicitly then the Default settings will be
applied.
4. If the computer that the user is logging in on has
an entry, then the computer settings are applied.
5. If there is not a computer entry for the user then
the default computer policy is applied.
Windows 95 policies are not compatible with NT and
therefore Windows 95 users must access a Windows 95
A long file name on a FAT partition uses one file name
for the 8.3 alias and then one more FAT entry for every
13 characters in the name.
A FAT partition can be converted to NTFS without
loss of data through the command line
CONVERT <drive>: /FS:NTFS
NTFS supports compression as a file attribute that can
be set in the file properties.
Compression can be applied to a folder or a drive and
the effect is that the files within are compressed and
any file copied into it will also become compressed.
Compression can be applied through the use of the
COMPACT.EXE program through the syntax
COMPACT <file or directory path> [/switch]
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 481
FAST FACTS: WINDOWS NT SERVER 4 EX AM
The available switches for COMPACT are as follows:
SUMMARY TABLE 9
COMPACT S W I T C H E S
Share-level permissions apply to any shared folder,
whether on FAT or NTFS.
NTFS permissions can only be applied to any file or
folder on an NTFS partition.
The actions that can be performed against an NTFS
object are as follows:
Switch
Function
/C
Compress
/U
Uncompress
á Read (R)
/S
Compress an entire directory tree
á Write (W)
/A
Compress hidden and system files
á Execute (X)
/I
Ignore errors and continue compressing
á Delete (D)
/F
Force compression even if the objects are already
compressed
á Change Permissions (P)
/Q
Display only summary information
á Take Ownership (O)
Share-level permissions apply only when users access a
resource over the network, not locally. The share-level
permissions are:
á No Access. Users with No Access to a share can
still connect to the share, but nothing appears in
File Manager except the message You do not have
permission to access this directory.
The NTFS permissions available for folders are summarized in the following table:
SUMMARY TABLE 10
NTFS FOLDER PERMISSIONS
Permission
Action permitted
No Access
none
á Read. Allows you to display folder and file
names, display file content and attributes, run
programs, open folders inside the shared folder.
List
RX
Read
RX
Add
WX
á Change. Allows you to create folders and files,
change file content, change file attributes, delete
files and folders, do everything READ permission
allows.
Add & Read
RXWD
Change
RXWD
Full Control
RXWDPO
á Full Control. Allows you to change file permissions and do everything change allows for.
Share-level permissions apply to the folder that is
shared and apply equally to all the contents of that
share.
481
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 482
482
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
The NTFS permissions available for files are summarized in the following table:
MONITORING AND
OPTIMIZATION
SUMMARY TABLE 11
Remember: Here are the elements that Microsoft says
they test on for the “Monitoring and Optimization”
section of the exam.
NT F S F I L E P E R M I S S I O N S
Permission
Action permitted
No Access
none
Read
RX
Add & Read
RX
á Monitor performance of various functions by
using Performance Monitor. Functions include:
processor, memory, disk, and network.
Change
RXWD
á Identify performance bottlenecks.
Full Control
RXWDPO
Performance monitor has 4 views: chart, alert, log, and
report.
If a user is given permission to a resource and a group
or groups that the user is a member is also given access
then the effective permission the user has is the cumulation of all of the user permissions. This applies unless
any of the permissions are set to No Access in which
case the user has no access to the resource.
The subsystems that are routinely monitored are:
Memory, Disk, Network, and Processor.
If a user is given permission to a shared resource and is
also given permission to that resource through NTFS
permissions then the effective permission is the most
restrictive permission.
Or
The File Child Delete scenario manifests itself when
someone has full control to a folder but is granted a
permission which does not enable deletion (Read or No
Access, for example). The effect is that a user will be
able to delete files inside the folder even though sufficient access does not appear to be present.
TROUBLESHOOTING
To close the File Child Delete loophole, do not grant a
user Full Control access to a folder but instead, use special Directory permissions to assign RXWDPO access,
this eliminates the File Child Delete permission.
Access Tokens do not refresh and a user needs to log off
and log back on if changed permissions are to take
effect.
Disk counters can be enabled through the command
line:
Diskperf –y
Diskperf –ye (for RAID disks and volumes)
Remember: Here are the elements that Microsoft says
they test on for the “Troubleshooting” section of the
exam.
á Choose the appropriate course of action to take
to resolve installation failures.
á Choose the appropriate course of action to take
to resolve boot failures.
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 483
FAST FACTS: WINDOWS NT SERVER 4 EX AM
483
á Choose the appropriate course of action to take
to resolve configuration errors.
The files involved in the boot process are identified in
the following table for both Intel and RISC machines:
á Choose the appropriate course of action to take
to resolve printer problems.
SUMMARY TABLE 12
á Choose the appropriate course of action to take
to resolve RAS problems.
F I L E S I N V O LV E D
Intel
RISC
á Choose the appropriate course of action to take
to resolve connectivity problems.
NTLDR
OSLOADER.EXE
BOOT.INI
NTOSKRNL.EXE
á Choose the appropriate course of action to take
to resolve fault tolerance problems. Fault-tolerance methods include: tape backup, mirroring,
stripe set with parity, and disk duplexing.
NTDETECT.COM
The acronym DETECT can be used to define the troubleshooting process and stands for:
á Discover the problem.
á Explore the boundaries.
IN THE
BOOT PROCESS
NTOSKRNL.EXE
In the NT boot process (in BOOT.INI) ARC paths
define the physical position of the NT operating system
files and come in two forms:
Scsi(0)disk(0)rdisk(0)partition(1)\WINNT
Multi(0)disk(0)rdisk(0)partition(1)\WINNT
á Track the possible approaches.
á Execute an Approach.
á Check for success.
á Tie up loose ends.
An NTHQ diskette can test a computer to ensure that
NT will successfully install on it.
The following list identifies possible sources of installation problems:
á Media errors
á Insufficient disk space
á Non-supported SCSI adapter
SCSI arc paths define hard drives which are SCSI and
which have their bios disabled. The relevant parameters are:
á SCSI: the SCSI controller starting from 0
á DISK: the physical disk starting from 0
á PARTITION: the partition on the disk stating
from 1
á \folder: the folder in which the NT files are
located
MULTI arc paths define hard drives which are nonSCSI or SCSI with their bios enabled. The relevant
parameters are:
á Failure of dependancy service to start
á MULTI: the controller starting from 0
á Inability to connect to the domain controller
á RDISK: the physical disk starting from 0
á Error in assigning domain name
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 484
484
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
á PARTITION: the partition on the disk stating
from 1
á \folder: the folder in which the NT files are
located
Partitions are numbered as follows:
Intel
RISC
BOOT.INI
*.PAL (for Alpha machines)
BOOTSECT.DOS (allows you to
boot to DOS)
NTBOOTDD.SYS (the SCSI driver for
a hard drive with SCSI bios not enabled)
1. The first primary partition on each disk gets the
number 0.
2. Each additional primary partition then is given a
number, incrementing up from 0.
3. Each logical drive is then given a number in the
order they appear in the Disk Administrator.
Switches on boot lines in the boot.ini file define additional boot parameters. The following table lists the
switches you need to know about and their function:
B O O T. I N I F I L E S W I T C H E S
Switch
Function
/basevideo
Loads standard VGA video driver (640x480,
16 color)
/sos
Displays each driver as it is loaded
/noserialmice
Prevents autodetection of serial mice on
COM ports which may disable a UPS connected to the port
A recovery disk can be used to bypass problems with
system partition. Such a disk contains the following
files (broken down by hardware platform):
SUMMARY TABLE 14
ON A
The RDISK programs allows you to update the
\REPAIR folder which in turn is used to update your
repair diskette.
The Event Viewer allows you to see three log files:
System Log, Security Log, and Application Log.
The Windows NT Diagnostics program allows you to
see (but not modify) configuration settings for much of
your hardware and environment.
SUMMARY TABLE 13
FILES
An Emergency repair disk can be used to recover an
NT system if the registry becomes corrupted and must
be used in conjunction with the three setup diskettes
used to install NT.
F A U LT - T O L E R A N T B O O T D I S K E T T E
Intel
RISC
NTLDR
OSLOADER.EXE
NTDETECT.COM
HAL.DLL
The course of action to take when a stop error occurs
(blue screen) can be configured from the System
Properties dialog box (in the Control Panel) on the
Startup/Shutdown tab.
To move the spool file from one partition to another,
use the Advanced Tab on the Server Properties dialog
box; this can be located from the File, Server Properties
menu in the printers dialog box.
Common RAS problems include the following:
á User Permission: user not enabled to use RAS in
User Manager for Domains.
á Authentication: often caused by incompatible
encryption methods (client using different
encryption than server is configured to receive).
á Callback with Multilink: Client configured for
callback but is using multilink; server will only
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 485
485
FAST FACTS: WINDOWS NT SERVER 4 EX AM
call back to a single number, thereby removing
multilink functionality.
á Autodial at Logon: Shortcuts on desktop referencing server-based applications or files causes
autodial to kick in when logon is complete.
User can’t login may be caused by a number of factors
including:
á Incorrect user name or password
á Incorrect domain name
á Incorrect user rights (inability to log on locally to
an NT machine, for example)
á Netlogon service on server is stopped or paused
á Domain controllers are down
á User is restricted in system policies from logging
on at a specific computer
The right to create backups and restore from backups
using NT Backup is granted to the groups Administrators, Backup Operators, and Server Operators by
default.
NT Backup will only backup files to tape, no other
media is supported.
The following table summarizes the backup types available in NT backup:
S U M M A R Y TA B L E 1 5
B A C K U P T Y P E S A VA I L A B L E
IN
NTBACKUP
Type
Backs Up
Marks?
Daily Copy
Selected files and folders
changed that day
No
The local registry of a computer can be backed up by
selecting the Backup Local Registry checkbox in the
Backup Information dialog box.
Data from tape can be restored to the original location
or to an alternate location and NTFS permissions can
be restored or not, however, you cannot change the
names of the objects being restored until the restore is
complete.
Backup can be run from a command line using the
NTBACKUP command in the syntax:
Ntbackup backup path [switches]
Some command line backup switches are shown in the
following table:
S U M M A R Y TA B L E 1 6
NTBACKUP COMMAND LINE SWITCHES
Switch
Function
/a
Append the current backup to the backup already on
the tape
/v
verify the backed up files when complete
/d “text”
Add an identifying description to the backup tape
/t option
specify the backup type. Valid options are: normal,
copy, incremental, differential, and daily
Type
Backs Up
Marks?
Normal
All selected files and folders
Yes
Copy
All selected files and folders
No
Incremental
Selected files and folders not
marked as backed up
Yes
1. Shut down your NT server and physically replace
the failed drive.
Differential
Selected files and folders not
marked as backed up
No
2. If required, boot NT using a recovery disk.
To recover from a failed mirror set you must do the following:
13 918-1 FastFact server.i 8/28/98 1:51 PM Page 486
486
FA S T FAC TS : W I ND OW S NT S E RVER 4 EXAM
3. Start the Disk Administrator using the menu
Start, Programs, Administrative Tools
(Common), Disk Administrator.
4. Select the mirror set by clicking on it.
5. From the Fault Tolerance menu choose Break
Mirror. This action exposes the remaining partition as a volume separate from the failed one.
6. Reestablish the mirror set if desired by selecting
the partition you desire to mirror and a portion
of free space equal in size and choosing the menu
Fault Tolerance, Establish Mirror.
To regenerate a stripe set with parity do the following:
1. Shut down your NT server and physically replace
the failed drive.
2. Start the Disk Administrator using the menu
Start, Programs, Administrative Tools
(Common), Disk Administrator.
3. Select the stripe set with parity by clicking on it.
4. Select an area of free space as large or larger than
the portion of the stripe set that was lost when
the disk failed.
5. Choose Fault Tolerance, Regenerate.
Hopefully, this has been a helpful tool in your final
review before the exam. You might find after reading
this that there are some places in the book you need to
revisit. Just remember to stay focused and answer all
the questions. You can always go back and check the
answers for the questions you are unsure of. Good luck!
14 918-1 Fast Facts ent.i 8/28/98 1:28 PM Page 487
The fast facts listed in this section are designed
as a refresher of key points and topics that are
required to succeed on the Windows NT server 4.0
in the Enterprise exam. By using these summaries
of key points, you can spend an hour prior to your
exam to refresh key topics, and ensure that you
have a solid understanding of the objectives and
information required for you to succeed in each
major area of the exam.
The following are the main categories Microsoft
uses to arrange the objectives:
á Planning
á Installation and configuration
á Managing resources
á Connectivity
á Monitoring and optimization
á Troubleshooting
For each of these main sections, or categories, the
assigned objectives are reviewed, and following
each objective, review material is offered.
PLANNING
Plan the implementation of a directory services architecture. Considerations include the following:
á Selecting the appropriate domain model
á Supporting a single logon account
á Enabling users to access resources in different
domains
Fast Facts
WINDOWS NT SERVER 4
ENTERPRISE EXAM
14 918-1 Fast Facts ent.i 8/28/98 1:28 PM Page 488
488
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
The main goals of directory services are the following:
á One user, one account
á Universal resource access
á Centralized administration
á Directory synchronization
To ensure that you are selecting the best plan for your
network, always address each of the goals of directory
services.
The requirements for setting up a trust are as follows:
á The trust relationship can be established only
between Windows NT Server domains.
á The domains must be able to make an RPC connection. To establish an RPC connection, you
must ensure that a network connection exists
between the domain controllers of all participating domains.
á The trust relationship must be set up by a user
with administrator access.
á You should determine the number and type of
trusts prior to the implementation.
á You must decide where to place the user
accounts, as that is the trusted domain.
Trust relationships enable communication between
domains. The trusts must be organized, however, to
achieve the original goal of directory services. Windows
NT domains can be organized into one of four different domain models:
á The single-domain model
á The single-master domain model
á The multiple-master domain model
á The complete-trust model
Table 1 summarizes the advantages and disadvantages
of the domain models.
TABLE 1
PROFILING
THE
DOMAIN MODELS
Domain Model
Advantages
Disadvantages
Single-domain
model
Centralized administration.
Limited to 40,000 user accounts.
No trust relationships.
No distribution of resources.
Single-master
domain model
Centralized administration.
Distributed resources.
Limited to 40,000 user accounts.
Multiple-master
domain model
Unlimited number of user
accounts; each master domain
can host 40,000 user accounts.
Distributed resources.
Complex trust relationships.
No centralized administration of user accounts.
Complete-trust
model
Unlimited number of user
accounts; each domain can
host 40,000 user accounts.
Complex trust relationships.
No centralized administration of user accounts.
14 918-1 Fast Facts ent.i 8/28/98 1:28 PM Page 489
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
Plan the disk drive configuration for various requirements. Requirements include choosing a fault-tolerance
method.
á Data Link Control (DLC)
á AppleTalk
Windows NT Server 4 comes bundled with several protocols that can be used for interconnectivity with other
systems and for use within a Windows NT environment. You examine the various protocols, then try to
define when each protocol best fits your network needs.
The protocols discussed to prepare you for the enterprise exam are the following:
Windows NT Server 4 supports the following faulttolerant solutions:
á RAID Level 0 (disk striping)
á RAID Level 1 (disk mirroring)
á RAID Level 5 (disk striping with parity)
A comparison of the three fault-tolerance options
might help to summarize the information and to ensure
that you have a strong understanding of the options
available in Windows NT Server 4 (see Table 2).
Choose a protocol for various situations. The protocols
include the following:
á TCP/IP
á TCP/IP with DHCP and WINS
á NWLink IPX/SPX Compatible Transport
Protocol
á NetBEUI. The NetBEUI protocol is the easiest
to implement and has wide support across platforms. The protocol uses NetBIOS broadcasts to
locate other computers on the network. This
process of locating other computers requires additional network traffic and can slow down your
entire network. Because NetBEUI uses broadcasts
to locate computers, it is not routable; in other
words, you cannot access computers that are not
on your physical network. Most Microsoft and
IBM OS/2 clients support this protocol.
NetBEUI is best suited to small networks with no
TABLE 2
SUMMARY
OF
F A U LT - T O L E R A N C E O P T I O N S
IN
WINDOWS NT SERVER 4
Disk Mirroring/
Disk Duplexing
Disk Striping with
Parity
No fault tolerance.
Complete disk
duplication.
Data regeneration from stored parity information.
Minimum of two
physical disks,
maximum of 32
disks.
Two physical disks
Minimum of three physical disks, maximum of 32 disks.
100 percent available
disk utilization.
50 percent available
disk utilization.
Dedicates the equivalent of one disk’s space in the set for parity information. The more
disks, the higher the utilization.
Cannot include a
system/boot partition.
Includes all partition
types.
Cannot include a system/boot partition.
Excellent read/write
performance.
Moderate read/write
performance.
Excellent read and moderate write performance.
Disk Striping
489
14 918-1 Fast Facts ent.i 8/28/98 1:28 PM Page 490
490
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
requirements for routing the information to
remote networks or to the Internet.
á TCP/IP. Transmission Control Protocol/Internet
Protocol, or TCP/IP, is the most common
protocol—more specifically, it is the most
common suite of protocols. TCP/IP is an
industry-standard protocol that is supported by
most network operating systems. Because of this
acceptance throughout the industry, TCP/IP
enables your Windows NT system to connect to
other systems with a common communication
protocol.
The following are advantages of using TCP/IP in
a Windows NT environment:
• The capability to connect dissimilar systems
• The capability to use numerous standard connectivity utilities, including File Transfer
Protocol (FTP), Telnet, and PING
• Access to the Internet
If your Windows NT system is using TCP/IP as a
connection protocol, it can communicate with
many non-Microsoft systems. Some of the systems it can communicate with are the following:
• Any Internet-connected system
• UNIX systems
• IBM mainframe systems
• DEC Pathworks
• TCP/IP-supported printers directly connected
to the network
á NWLink IPX/SPX Compatible. The IPX
protocol has been used within the NetWare
environment for years. By developing an IPXcompatible protocol, Microsoft enables Windows
NT systems to communicate with NetWare
systems.
NWLink is best suited to networks requiring
communication with existing NetWare servers
and for existing NetWare clients.
Other utilities must be installed, however, to
enable the Windows NT Server system to gain
access into the NetWare security. Gateway
Services for NetWare/Client Services for NetWare
(GSNW/CSNW) must be installed on the
Windows NT server to enable the computer to
be logged on to a NetWare system. GSNW functions as a NetWare client, but it also can share
the connection to the Novell box with users from
the Windows NT system. This capability enables
a controlled NetWare connection for file and
print sharing on the NetWare box, without
requiring the configuration of each NT client
with a duplicate network redirector or client.
á DataLink Control. The DLC protocol was originally used for connectivity in an IBM mainframe
environment, and maintains support for existing
legacy systems and mainframes. The DLC protocol is also used for connections to some network
printers.
á AppleTalk. Windows NT Server can configure
the AppleTalk protocol to enable connectivity
with Apple Macintosh systems. This protocol is
installed with the Services for the Macintosh
included with your Windows NT Server
CD-ROM. The AppleTalk protocol enables
Macintosh computers on your network to access
files and printers set up on the Windows NT
server. It also enables your Windows NT clients
to print to Apple Macintosh printers.
The AppleTalk protocol is best suited to connectivity with the Apple Macintosh.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 491
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
INSTALLATION AND
CONFIGURATION
Install Windows NT Server to perform various server
roles. Server roles include the following:
á Primary domain controller
á Backup domain controller
á Member server
The following are different server roles into which
Windows NT Server can be installed:
á Primary Domain Controller. The Primary
Domain Controller (PDC) is the first domain
controller installed into a domain. As the first
computer in the domain, the PDC creates the
domain. This fact is important to understand
because it establishes the rationale for needing a
PDC in the environment. Each domain can contain only one PDC. All other domain controllers
in the domain are installed as Backup Domain
Controllers. The PDC handles user requests and
logon validation, and it offers all the standard
Windows NT Server functionality. The PDC
contains the original copy of the Security
Accounts Manager (SAM), which contains all
user accounts and security permissions for your
domain.
á Backup Domain Controller. The Backup
Domain Controller (BDC) is an additional
domain controller used to handle logon requests
by users in the network. To handle the logon
requests, the BDC must have a complete copy of
the domain database, or SAM. The BDC also
runs the Netlogon service; however, the Netlogon
service in a BDC functions a little differently
than in a PDC. In the PDC, the Netlogon
491
service handles synchronization of the SAM database to all the BDCs.
á Member server. In both of the domain controllers, PDC or BDC, the computer has an additional function: The domain controllers handle
logon requests and ensure that the SAM is synchronized throughout the domain. These functions add overhead to the system. A computer
that handles the server functionality you require
without the overhead of handling logon validation is called a member server. A member server is
a part of the domain, but it does not need a copy
of the SAM database and does not handle logon
requests. The main function of a member server
is to share resources.
After you have installed your computer into a specific
server role, you might decide to change the role of the
server. This can be a relatively easy task if you are
changing a PDC to a BDC or vice versa. If you want to
change a domain controller to a member server or
member server to a domain controller, however, you
must reinstall into the required server role. A member
server has a local database that does not participate in
domain synchronization. In changing roles, a member
server must be reinstalled to ensure that the account
database and the appropriate services are installed.
Configure protocols and protocol bindings. Protocols
include the following:
á TCP/IP
á TCP/IP with DHCP and WINS
á NWLink IPX/SPX Compatible Transport
Protocol
á DLC
á AppleTalk
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 492
492
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
NOTE
You install a new protocol in Windows NT Server
through the Network Properties dialog box.
NetBEUI Not Discussed This list
does not include the NetBEUI protocol, as there are no configuration
options available for this protocol.
Following are the protocols, and the configuration
options available with each:
• DNS. The DNS tab shows you the options
available for configuring your TCP/IP protocol to use a DNS server. The Domain Name
System (DNS) server translates TCP/IP host
names of remote computers into IP addresses.
Remember that an IP address is a unique
address for each computer. The DNS server
contains a database of all the computers you
can access by host name. This database is used
when you access a Web page on the Internet.
Working with the naming scheme is easier
than using the IP address of the computer.
• IP Address. The IP Address tab enables you
to configure the IP address, the subnet mask,
and the default gateway. You also can enable
the system to allocate IP address information
automatically through the use of the DHCP
server.
• WINS Address. The WINS Address tab
enables you to configure your primary and
secondary Windows Internet Names Services
(WINS) server addresses. WINS is used to
reduce the number of NetBIOS broadcast
messages sent across the network to locate a
computer. By using a WINS server, you keep
the names of computers on your network in a
WINS database. The WINS database is
dynamic.
An IP address is a 32-bit address that is broken into four octets and used to identify your
network adapter card as a TCP/IP host. Each
IP address must be a unique address. If you
have any IP address conflicts on your computer, you cannot use the TCP/IP protocol.
In configuring your WINS servers, you can
enter your primary WINS server and a secondary WINS server. Your system searches the
primary WINS server database first, then the
secondary database if no match was found in
the primary one.
Your IP address is then grouped into a subnet.
The process you use to subnet your network
is to assign a subnet mask. A subnet mask is
used to identify the computers local to your
network. Any address outside your subnet is
accessed through the default gateway, also
called the router. The default gateway is the
address of the router that handles all routing
of your TCP/IP information to computers, or
hosts, outside your subnet.
• DHCP Relay. The DHCP relay agent is used
to find your DHCP servers across routers.
DHCP addresses are handed out by the
DHCP servers. The client request, however, is
made with a broadcast message. Broadcast
messages do not cross routers; therefore, this
protocol might place some restrictions on
your systems. The solution is to use a DHCP
relay agent to assist the clients in finding the
DHCP server across a router.
á TCP/IP. The following tabs are available for configuration in the Microsoft TCP/IP Properties
dialog box:
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 493
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
In configuring your DHCP relay agent, you
can specify the seconds threshold and the
maximum number of hops to use in searching
for the DHCP servers. At the bottom of the
tab, you can enter the IP addresses of the
DHCP servers you want to use.
• Routing. In an environment in which multiple subnets are used, you can configure your
Windows NT Server as a multihomed system.
In other words, you can install multiple network adapters, each connecting to a different
subnet. If you enable the Enable IP
Forwarding option, your computer acts as a
router, forwarding the packets through the
network cards in the multihomed system to
the other subnet.
á NWLINK IPX/SPX Compatible. The
configuration of the NWLink protocol is simple
in comparison to the TCP/IP protocol. It is this
simplicity that makes it a popular protocol to use.
The NWLink IPX/SPX Properties dialog box has
two tabs:
• General. On the General tab, you have the
option to assign an internal network number.
This eight-digit hexadecimal number format
is used by some programs with services that
can be accessed by NetWare clients.
You also have the option to select a frame
type for your NWLink protocol. The frame
type you select must match the frame type of
the remote computer with which you need to
communicate. By default, Windows NT
Server uses the Auto Frame Type Detection
setting, which scans the network and loads
the first frame type it encounters.
• Routing. The Routing tab of the NWLink
IPX/SPX Properties dialog box is used to
493
enable or disable the Routing Information
Protocol (RIP). If you enable RIP routing
over IPX, your Windows NT Server can act
as an IPX router.
á DLC. The configuration of DLC is done through
Registry parameters. The DLC protocol is configured based on three timers:
• T1. The response timer
• T2. The acknowledgment delay timer
• Ti. The inactivity timer
The Registry contains the entries that can be
modified to configure DLC. You can find the
entries at
HKEY_LOCAL_MACHINE\SYSTEM\Current
ControlSet\Services\DLC\Parameters\ELNKIII
adapter name
á AppleTalk. To install the AppleTalk protocol,
you install Services for Macintosh.
Table 3 reviews the protocols that you can configure for
your NT enterprise (including the subcomponents—
tabs—of each protocol).
TABLE 3
PROTOCOLS
TO
CONFIGURE
Protocol
Subcomponent (Tab)
TCP/IP
IP Address
DNS
WINS Address
DHCP Relay
Routing
NWLink IPX/SPX Compatible
General
Routing
AppleTalk
General
Routing
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 494
494
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
The binding order is the sequence your computer uses
to select which protocol to use for network communications. Each protocol is listed for each network-based
service, protocol, and adapter available.
The Bindings tab contains an option, Show Bindings
for, that can be used to select the service, adapter, or
protocol you want to modify in the binding order. By
clicking the appropriate button, you can enable or disable each binding, or move up or down in the binding
order.
Configure Windows NT Server core services. Services
include the following:
á Directory Replicator
á Computer Browser
In this objective, you look at configuring some of the
core services in the Windows NT Server. These services
are the following:
á Server service. The Server service answers network requests. By configuring Server service, you
can change the way your server responds and, in
a sense, the role it plays in your network environment. To configure Server service, you must open
the Network dialog box. To do this, double-click
the Network icon in the Control Panel. Select the
Services tab. In the Server dialog box, you have
four optimization settings. Each of these settings
modifies memory management based on the role
the server is playing. These options are the following:
• Minimize Memory Used. The Minimize
Memory Used setting is used when your
Windows NT Server system is accessed by less
than 10 users.
This setting allocates memory so a maximum
of 10 network connections can be properly
maintained. By restricting the memory for
network connections, you make more memory available at the local or desktop level.
• Balance. The Balance setting can be used for
a maximum of 64 network connections. This
setting is the default when using NetBEUI
software. Like the Minimize setting, Balance
is best used for a relatively low number of
users connecting to a server that also can be
used as a desktop computer.
• Maximize Throughput for File Sharing.
The Maximize Throughput for File Sharing
setting allocates the maximum amount of
memory available for network connections.
This setting is excellent for large networks in
which the server is being accessed for file and
print sharing.
• Maximize Throughput for Network
Applications. If you are running distributed
applications, such as SQL Server or Exchange
Server, the network applications do their own
memory caching. Therefore, you want your
system to enable the applications to manage
the memory. You accomplish this by using the
Maximize Throughput for Network
Applications setting. This setting also is used
for very large networks.
á Computer Browser service. The Computer
Browser service is responsible for maintaining the
list of computers on the network. The browse list
contains all the computers located on the physical
network. As a Windows NT Server, your system
plays a big role in the browsing of a network. The
Windows NT Server acts as a master browser or
backup browser.
The selection of browsers is through an election.
The election is called by any client computer or
when a preferred master browser computer starts
up. The election is based on broadcast messages.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 495
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
Every computer has the opportunity to nominate
itself, and the computer with the highest settings
wins the election.
The election criteria are based on three things:
• The operating system (Windows NT Server,
Windows NT Workstation, Windows 95,
Windows for Workgroups)
• The version of the operating system (NT 4.0,
NT 3.51, NT 3.5)
• The current role of the computer (master
browser, backup browser, potential browser)
á Directory Replicator service. You can configure
the Directory Replicator service to synchronize an
entire directory structure across multiple servers.
In configuring the directory service, you must
select the export server and all the import servers.
The export server is the computer that holds the
original copy of the directory structure and files.
Each import server receives a complete copy of
the export server’s directory structure. The
Directory Replicator service monitors the directory structure on the export server. If the contents
of the directory change, the changes are copied to
all the import servers. The file copying and directory monitoring is completed by a special service
account you create. You must configure the
Directory Replicator service to use this service
account. The following access is required for your
Directory Replicator service account:
• The account should be a member of the
Backup Operators and Replicators groups.
• There should be no time or logon restrictions
for the account.
• The Password Never Expires option should be
selected.
495
• The User Must Change Password At Next
Logon option should be turned off.
When configuring the export server, you have the
option to specify the export directory. The default
export directory is
C:\WINNT\system32\repl\export\.
In the Import Directories section of the Directory
Replication dialog box, you can select the import
directory. The default import directory is
C:\WINNT\system32\repl\import.
Remember that the default directory for executing logon scripts in a Windows NT system is
C:\WINNT\system32\repl\import\scripts.
Configure hard disks to meet various requirements.
Requirements include the following:
á Providing duplication
á Improving performance
All hard disk configuration can be done using the Disk
Administrator tool. The different disk configurations
you need to understand for the enterprise exam are the
following:
á Stripe set. A stripe set gives you improved disk
read and write performance; however, it supplies
no fault tolerance. A minimum of two disks is
required, and the configuration can stripe up to
32 physical disks. A stripe set cannot include the
system partition.
á Volume set. A volume set enables you to extend
partitions beyond one physical disk; however, it
supplies no fault tolerance. To extend a volume
set, you must use the NTFS file system.
á Disk mirroring. A mirror set uses two physical
disks and provides full data duplication. Often
referred to as RAID level 1, disk mirroring is a
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 496
496
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
useful solution to assigning duplication to the
system partition, as well as any other disks that
might be in the system.
á Stripe set with parity. A stripe set with parity
enables fault tolerance in your system. A minimum of three physical disks is required, and a
maximum of 32 physical disks can be included in
a stripe set with parity. A stripe set with parity
cannot include the system partition of your
Windows NT system.
The solution that supplies the best duplication and
optimization mix is the stripe set with parity.
Configure printers. Tasks include the following:
á Adding and configuring a printer
á Implementing a printer pool
á Setting print priorities
The installation of a printer is a fairly simplistic procedure and is not tested heavily on the exam; however,
the printer pool is a key point. The items to remember
about printer pools are as follows:
á All printers in a printer pool must be able to
function using the same printer driver.
á A printer pool can have a maximum of eight
printers in the pool.
Configure a Windows NT Server computer for various
types of client computers. Client computer types
include the following:
á Windows NT Workstation
á Windows 95
á Macintosh
The Network Client Administrator is found in the
Administrative Tools group. You can use the Network
Client Administrator program to do the following:
á Make a Network Installation Startup Disk.
This option creates an MS-DOS boot disk that
contains commands required to connect to a network server and that automatically installs
Windows NT Workstation, Windows 95, or the
DOS network clients.
á Make an Installation Disk Set. This option
enables the creation of installation disks for the
DOS network client, LAN Manager 2.2c for
DOS, or LAN Manager 2.2c for OS/2.
á Copy Client-Based Network Administration
Tools. This option enables you to share the network administration tools with client computers.
The client computers that can use the network
administration tools are Windows NT
Workstation and Windows 95 computers.
á View Remoteboot Client Information. This
option enables you to view the remoteboot client
information. To install remoteboot, go to the
Services tab of the Network dialog box.
When installing a client computer, you must ensure
that your Windows NT system is prepared for and configured for the client. The Windows clients can connect
to the Windows NT server without any configuration
required on the server; however, some configuration is
required on the client computers. For the Apple
Macintosh client, the NT server must install the services for the Macintosh, which includes the AppleTalk
protocol. This protocol enables the seamless connection
between the Windows NT system and the Apple
clients.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 497
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
MANAGING RESOURCES
Manage user and group accounts. Considerations
include the following:
á Managing Windows NT user accounts
á Managing Windows NT user rights
á Managing Windows NT groups
á Administering account policies
á Auditing changes to the user account database
AGLP stands for Accounts/Global Groups/Local
Groups/Permissions. When you want to assign permissions to any resource, you should follow a few simple
rules. All user accounts are placed into global groups,
and global groups get assigned into local groups. The
local groups have the resources and permissions
assigned to them.
When you are working with groups across trust relationships, the following guidelines are useful:
á Always gather users into global groups.
Remember that global groups can contain only
user accounts from the same domain. You might
have to create the same named global group in
multiple domains.
á If you have multiple account domains, use the
same name for a global group that has the same
types of members. Remember that when multiple
domains are involved, the group name is referred
to as DOMAIN\GROUP.
á Before the global groups are created, determine
whether an existing local group meets your needs.
There is no sense in creating duplicate local
groups.
á Remember that the local group must be created
where the resource is located. If the resource is on
497
a Domain Controller, create the local group in
the Domain Account Database. If the resource is
on a Windows NT Workstation or Windows NT
Member Server, you must create the group in
that system’s local account database.
á Be sure to set the permissions for a resource
before you make the global groups a member of
the local group assigned to the resource. That
way, you set the security for the resource.
Create and manage policies and profiles for various situations. Policies and profiles include the following:
á Local user profiles
á Roaming user profiles
á System policies
You can configure system policies to do the following:
á Implement defaults for hardware configuration—
for all computers using the profile or for a specific
machine.
á Restrict the changing of specific parameters that
affect the hardware configuration of the participating system.
á Set defaults for all users in the areas of their personal settings that the users can configure.
á Restrict users from changing specific areas of their
configuration to prevent tampering with the system. An example is disabling all Registry editing
tools for a specific user.
á Apply all defaults and restrictions on a group
level rather than just a user level.
Some common implementations of user profiles are the
following:
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 498
498
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
á Locking down display properties to prevent users
from changing the resolution of their monitor.
Display properties can be locked down as a whole
or on each individual property page of display
properties. You adjust this setting by clicking the
Control Panel, Display, Restrict Display option of
the Default User Properties dialog box.
á Setting a default color scheme or wallpaper. You
can do this by clicking the Desktop option of the
Default User Properties dialog box.
á If you want to restrict access to portions of the
Start menu or desktop, you can do this by clicking the Shell, Restrictions option of the Default
User Properties dialog box.
á If you need to limit the applications that the user
can run at a workstation, you can do so by clicking the System, Restrictions option of the Default
User Properties dialog box. You can also use this
option to prevent the user from modifying the
Registry.
á You can prevent users from mapping or disconnecting network drives by clicking the Windows
NT Shell, Restrictions option of the Default User
Properties dialog box.
Profiles and policies can be very powerful tools to assist
in the administrative tasks in your environment. The
following list reviews each of the main topics covered in
this objective:
á Roaming profiles. The user portion of the
Registry is downloaded from a central location,
allowing the user settings to follow the user anywhere within the network environment.
á Local profiles. The user settings are stored at
each workstation and are not copied to other
computers. Each workstation that you use will
have different desktop and user settings.
á System policies. System policies enable the
administrator to restrict user configuration
changes on systems. This enables the administrator to maintain the settings of the desktop of systems without the fear that a user can modify
them.
á Computer policies. Computer policies allow the
lockdown of common machine settings that
affect all users of that computer.
Administer remote servers from various types of client
computers. Client computer types include the following:
á Windows 95
á Windows NT Workstation
This objective focuses on the remote administration
tools available for your Windows NT Server. The following list summarizes the key tools:
á Remote Administration Tools for Windows 95.
Allows User Manager, Server Manager, Event
Viewer, and NTFS file permissions to be executed from the Windows 95 computer.
á Remote Administration for Windows NT.
Allows User Manager, Server Manager, DHCP
Manager, System Policy Editor, Remote Access
Admin, Remote Boot Manager, WINS Manager,
and NTFS file permissions to be executed from a
Windows NT machine.
á Web Based Administration. Allows for common
tasks to be completed through an Internet connection into the Windows NT Server.
Manage disk resources. Tasks include the following:
á Creating and sharing resources
á Implementing permissions and security
á Establishing file auditing
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 499
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
Windows NT has two levels of security for protecting
your disk resources:
á Share permissions
á NTFS permissions
NTFS permissions enable you to assign more
comprehensive security to your computer system.
NTFS permissions can protect you at the file level.
Share permissions, on the other hand, can be applied
only to the folder level. NTFS permissions can affect
users logged on locally or across the network to the
system where the NTFS permissions are applied. Share
permissions are in effect only when the user connects to
the resource through the network.
The combination of Windows NT share permissions
and NTFS permissions determines the ultimate access a
user has to a resource on the server’s disk. When share
permissions and NTFS permissions are combined, no
preference is given to one or the other. The key factor is
which of the two effective permissions is the most
restrictive.
For the exam, remember the following tips relating to
managing resources:
á Users can be assigned only to global groups in the
same domain.
á Only global groups from trusted domains can
become members of local groups in trusting
domains.
á NTFS permissions are assigned only to local
groups in all correct test answers.
á Only NTFS permissions give you file-level
security.
499
CONNECTIVITY
Configure Windows NT Server for interoperability
with NetWare servers by using various tools. The tools
include the following:
á Gateway Service for NetWare
á Migration Tool for NetWare
Gateway Service for NetWare (GSNW) performs the
following functions:
á GSNW enables Windows NT Servers to access
NetWare file and print resources.
á GSNW enables the Windows NT Servers to act
as a gateway to the NetWare file and print
resources. The Windows NT Server enables users
to borrow the connection to the NetWare server
by setting it up as a shared connection.
The Migration Tool for NetWare (NWCONV) transfers file and folder information and user and group
account information from a NetWare server to a
Windows NT domain controller. The Migration Tool
can preserve the folder and file permissions if it is being
transferred to an NTFS partition.
Connectivity between Windows NT and a NetWare
server requires the use of GSNW. If the user and file
information from NetWare is to be transferred to a
Windows NT Server, the NetWare Conversion utility,
NWCONV, is used for this task. The following list
summarizes the main points in this section on NetWare
connectivity:
á GSNW can be used as a gateway between
Windows NT clients and a NetWare server.
á GSNW acts as a NetWare client to the Windows
NT Server, allowing the NT server to have a connection to the NetWare server.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 500
500
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
á GSNW is a service in Windows NT, and is
installed using the Control Panel.
á For GSNW to be used as a gateway into a
NetWare server, a gateway user account must be
created and placed in a NetWare group called
NTGATEWAY.
á In configuring the GSNW as a gateway, you can
assign permissions to the gateway share by accessing the GSNW icon in the Control Panel.
á For GSNW to be functional, the NWLINK
IPX/SPX protocol must be installed and
configured.
á To convert user and file information from a
NetWare server to a Windows NT server, you can
use the NWCONV.EXE utility.
á NWCONV requires that GSNW be installed
prior to any conversion being carried out.
á To maintain the NetWare folder- and file-level
permissions in the NWCONV utility, you must
convert to an NTFS partition on the Windows
NT system.
Install and configure multiprotocol routing to serve
various functions. Functions include the following:
á Internet router
á BOOTP/DHCP Relay Agent
á IPX router
Multiprotocol routing gives you flexibility in the connection method used by your clients, and in maintaining security. Check out the following:
á Internet router. Setting up Windows NT as an
Internet router is as simple as installing two network adapters in the system, then enabling IP
routing in the TCP/IP protocol configuration.
This option enables Windows NT to act as a
static router. Note that Windows NT cannot
exchange Routing Information Protocol (RIP)
routing packets with other IP RIP routers unless
the RIP routing software is installed.
á IPX router. You enable the IPX router by
installing the IPX RIP router software by choosing Control Panel, Networks, Services.
After installing the IPX RIP router, Windows NT
can route IPX packets over the network adapters
installed. Windows NT uses the RIP to exchange
its routing table information with other RIP
routers.
The inclusion of the industry-standard protocols, and
tools to simplify the configuration and extension of
your NT network into other environments, makes this
operating system a very powerful piece of your heterogenous environment. The following are the main
factors to focus on for this objective:
á A strong understanding of the functionality of
each of the Windows NT protocols—with a
strong slant toward TCP/IP and the configuration options available. Understanding and configuration of the DHCP server are also tested on
this exam.
á The services used to resolve the IP addresses and
names of hosts in a TCP/IP environment. DNS
service, WINS Service, the Hosts file, and the
LMHosts files are among the services tested.
á The routing mechanisms available in Windows
NT. These mechanisms are powerful, and largely
unknown to the vast majority of NT administrators. Ensure that you review the configuration
and functionality of Internet or IP routing, as
well as the IPX routing tools available.
Install and configure Internet Information Server, and
install and configure Internet services. Services include
the following:
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 501
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
á The World Wide Web
á DNS
á Intranets
501
Install and configure Remote Access Service (RAS).
Configuration options include the following:
á Configuring RAS communications
á Configuring RAS protocols
Internet Information Server (IIS) uses Hypertext
Transfer Protocol (HTTP), File Transfer Protocol
(FTP), and the Gopher service to provide Internet publishing services to your Windows NT Server computer.
IIS provides a graphical administration tool called the
Internet Service Manager. With this tool, you can centrally manage, control, and monitor the Internet services in your Windows NT network. The Internet
Service Manager uses the built-in Windows NT security model, so it offers a secure method of remotely
administering your Web sites and other Internet services.
IIS is an integrated component in Windows NT Server
4.0. The IIS services are installed using the Control
Panel, Networks icon or during the installation phase.
The following list summarizes the key points in
installing and configuring IIS:
á The three Internet services included in IIS are
HTTP, FTP, and Gopher.
á HTTP is used to host Web pages from your
Windows NT server system.
á FTP is a protocol used for transferring files across
the Internet using the TCP/IP protocol.
á Gopher is used to create a set of hierarchical links
to other computers or to annotate files or folders.
á The Internet Service Manager is the utility used
to manage and configure your Internet services in
IIS.
á The Internet Service Manager has three views
that you can use to view your services. The three
views are Report View, Servers View, and Services
View.
á Configuring RAS security
RAS supports the Serial Line Internet Protocol (SLIP)
and Point-to-Point Protocol (PPP) line protocols, and
the NetBEUI, TCP/IP, and IPX network protocols.
RAS can connect to a remote computer using any of
the following media:
á Public Switched Telephone Network (PSTN).
(PSTN is also known simply as the phone company.) RAS can connect using a modem through
an ordinary phone line.
á X.25. A packet-switched network. Computers
access the network through a Packet Assembler
Disassembler (PAD) device. X.25 supports dialup or direct connections.
á Null modem cable. A cable that connects two
computers directly. The computers then communicate using their modems (rather than network
adapter cards).
á ISDN. A digital line that provides faster communication and more bandwidth than a normal
phone line. (It also costs more, which is why not
everybody has it.) A computer must have a special ISDN card to access an ISDN line.
RAS is designed for security. The following are some of
RAS’s security features:
á Auditing. RAS can leave an audit trail, enabling
you to see who logged on when and what authentication they provided.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 502
502
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
á Callback security. You can enable the RAS server
to use callback (hang up all incoming calls and
call the caller back), and you can limit callback
numbers to prearranged sites that you know are
safe.
á Encryption. RAS can encrypt logon information,
or it can encrypt all data crossing the connection.
á Security hosts. In case Windows NT is not safe
enough, you can add an extra dose of security by
using a third-party intermediary security host—a
computer that stands between the RAS client and
the RAS server and requires an extra round of
authentication.
á PPTP filtering. You can tell Windows NT to filter out all packets except ultra safe Point-to-Point
Tunneling Protocol (PPTP) packets.
RAS can be a very powerful and useful tool in enabling
you to extend the reaches of your network to remote
and traveling users. The following list summarizes main
points for RAS in preparation for the exam:
á RAS supports SLIP and PPP line protocols.
á With PPP, RAS can support NetBEUI,
NWLINK, and TCP/IP across the communication line.
á RAS uses the following media to communicate
with remote systems: PSTN, X.25, Null Modem
cable, and ISDN.
á The RAS security features available are auditing,
callback security, encryption, and PPTP filtering.
á To install RAS, click the Network icon in the
Control Panel.
MONITORING AND
OPTIMIZATION
Establish a baseline for measuring system performance.
Tasks include creating a database of measurement data.
You can use numerous database utilities to analyze the
data collected. The following are some of the databases
that Microsoft provides:
á Performance Monitor
á Microsoft Excel
á Microsoft Access
á Microsoft FoxPro
á Microsoft SQL Server
The following list summarizes the key items to focus
on when you are analyzing your computer and network:
á Establish a baseline measurement of your system
when functioning at its normal level. Later, you
can use the baseline in comparative analysis.
á Establish a database to maintain the baseline
results and any subsequent analysis results on the
system, to compare trends and identify potential
pitfalls in your system.
á The main resources to monitor are memory, the
processor, the disks, and the network.
The following list summarizes the tools used to monitor your NT server that are available and are built into
Windows NT Server 4.0:
á Server Manager
á Windows NT Diagnostics
á Response Probe
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 503
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
á Performance Monitor
á Network Monitor
Monitor performance of various functions by using
Performance Monitor. Functions include the following:
á Processor
á Memory
á Disk
á Network
To summarize the main views used within Performance
Monitor, review the following list:
á Chart view. This view is very useful for viewing
the objects and counters in a real-time mode.
This mode enables you to view the data in a
graphical format. You can also use the chart view
to view the contents of a log file.
á Log view. This view enables you to set all the
options required for creating a log of your system
resources or objects. After this log is created, you
can view it by using the chart view.
á Alert view. Use the alert view to configure warnings or alerts of your system resources or objects.
In this view, you can configure threshold levels
for counters and can then launch an action based
on the threshold values being exceeded.
á Report view. The report view enables you to
view the object and counters as an averaged
value. This view is useful for comparing the values of multiple systems that are configured similarly.
When you want to monitor TCP/IP counters, make
sure that SNMP is installed. Without the SNMP service
installed, the TCP/IP counters are not available.
Performance Monitor is a graphical utility that you can
use for monitoring and analyzing your system resources
within Windows NT. You can enable objects and counters within Performance Monitor; it is these elements
that enable the logging and viewing of system data.
In preparing you for this objective, this section introduces numerous objects and counters that you use with
Performance Monitor. To prepare for the exam, you
need to understand the following key topics:
á The four views available in Performance Monitor
are the report view, the log view, the chart view,
and the alert view.
á The main resources to monitor in any system are
the disk, the memory, the network, and the
processor.
á Each of the main resources is grouped as a
separate object, and within each object are
counters. A counter is the type of data available
from a type of resource or object. Each counter
might also have multiple instances. An instance is
available if multiple components in a counter are
listed.
á To enable the disk counters to be active, you
must run the DISKPERF utility.
Monitor network traffic by using Network Monitor.
Tasks include the following:
á Collecting data
á Presenting data
á Filtering data
When monitoring the disk, remember to activate the
disk counters using the command diskperf –y. If you do
not enter this command, you can select counter but will
not see any activity displayed. In the case of a software
RAID system, start diskperf with the -ye option.
503
Network Monitor is a network packet analyzer that
comes with Windows NT Server 4. Actually, two
versions of Network Monitor are available from
Microsoft.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 504
504
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
The first version comes with Windows NT Server 4
(simple version). This version can monitor the packets
(frames) sent or received by a Windows NT Server 4
computer. The second version comes with Microsoft
Systems Management Server (full version). This version
can monitor all traffic on the network.
Identify performance bottlenecks and optimize
performance for various results. Results include the
following:
By fully understanding the various components found
while analyzing traffic, you will be more successful in
locating potential network bottlenecks and offering relevant optimization recommendations. The main components that need to be monitored with your network
traffic analysis are the following:
To optimize the logon traffic in your Windows NT
network, you should consider four main points:
á Locate and classify each service. Analyze the
amount of traffic generated from each individual
service, the frequency of the traffic, and the overall effect the traffic has on the network segment.
á Understand the three different types of frames:
broadcast, multicast, and directed.
á Review the contents of a frame and ensure that
you can find the destination address, source
address, and data located in each frame.
The following points summarize the key items to
understand in building a strong level of knowledge in
using Network Monitor as a monitoring tool:
á Two versions of Network Monitor are available:
the scaled-down version that is built into the
Windows NT Server operating system, and the
full version that is a component of Microsoft
Systems Management Server.
á The Network Monitor windows consist of four
sections: Graph, Session Statistics, Station
Statistics, and Total Statistics.
á After Network Monitor captures some data, you
use the display window of Network Monitor to
view the frames. The three sections of the display
window are the Summary pane, the Detail pane,
and the Hexadecimal pane.
á Controlling network traffic
á Controlling the server load
á Determine the hardware required to increase performance.
á Configure the domain controllers to increase the
number of logon validations.
á Determine the number of domain controllers
needed.
á Determine the best location for each of the
domain controllers.
The following are a few good points to follow in optimizing file-session traffic:
á Remove any excess protocols that are loaded.
á Reduce the number of wide area network (WAN)
links required for file transfer.
The following are three points to consider when
attempting to optimize server browser traffic:
á Reduce the number of protocols.
á Reduce the number of entries in the browse list.
á Increase the amount of time between browser
updates.
Trust relationships generate a large amount of network
traffic. In optimizing your system, attempt to keep the
number of trusts very low.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 505
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
TROUBLESHOOTING
Choose the appropriate course of action to take to
resolve installation failures.
Troubleshooting a Windows NT system requires that
you have a strong understanding of the processes and
tools available to you. To be an effective troubleshooter,
first and foremost you must have experience. The following is a list of some common installation problems:
á Hard disk problems
á Unsupported CD-ROMs
á Network adapter problems and conflicts
á Naming problems (each computer must be
uniquely named, following the NetBIOS naming
conventions)
Always use the hardware compatibility list to ensure
that your components are supported by Windows NT.
Choose the appropriate course of action to take to
resolve boot failures.
For startup errors, try the following:
á Check for missing files that are involved in the
boot process, including NTLDR, NTDETECT.COM, BOOT.INI, NTOSKRNL.EXE,
and OSLOADER (RISC).
505
You can resolve many problems that you encounter
within Windows NT by configuring the Registry.
However, before you make any Registry configurations,
you must have a strong understanding of the keys within the Registry and always back up the Registry prior to
making any modifications to ensure a smooth rollback
if additional problems occur. The following are the
main tools used to modify the Registry:
á REGEDT32
á REGEDIT
For configuration problems, remember the following:
á Using the Registry for configuration and troubleshooting can cause additional problems if you
do not maintain a full understanding of the
Registry.
á Always back up the Registry prior to editing the
contents.
á You can back up and restore the local Registry by
using REGEDT32.
Choose the appropriate course of action to take to
resolve printer problems.
For troubleshooting printers, you should do the following:
á Modify BOOT.INI for options.
á Understand and review the overview of the printing process.
á Create an NT boot disk for bypassing the boot
process from the hard disk.
á Understand the files involved in the printing
process.
á Use the Last Known Good option to roll back to
the last working set of your Registry settings.
á As a first step in troubleshooting a printer, always
verify that the printer is turned on and online.
Choose the appropriate course of action to take to
resolve configuration errors. Tasks include the following:
á Backing up and restoring the Registry
á Editing the Registry
á Note that the most common errors associated
with a printer are an invalid printer driver or
incorrect resource permissions set for a user.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 506
506
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
Choose the appropriate course of action to take to
resolve RAS problems.
The following is a list of some of the problems that you
might encounter with RAS:
á You must ensure that the protocol you are
requesting from the RAS client is available on the
RAS server. There must be at least one common
protocol or the connection will fail.
á If you are using NetBEUI, ensure that the name
you are using on the RAS client is not in use on
the network to which you are attempting to connect.
á If you are attempting to connect using TCP/IP,
you must configure the RAS server to provide
you with an address.
You can use the Remote Access Admin tool to monitor
the ports as well as the active connections of your RAS
server.
Numerous RAS settings can cause some problems with
your RAS connections. Ensure that you understand the
installation process, as well as any configuration settings
required to enable your RAS server. You can avoid
some of the common problems that can occur by doing
the following:
á Ensuring that the modem and communication
medium are configured and functional prior to
installing RAS. It can be very difficult to modify
settings after the installation, so it is recommended to have all hardware tested and working first.
á Verifying that dial-in permissions have been
enabled for the required users. This small task is
commonly forgotten in your RAS configuration.
Choose the appropriate course of action to take to
resolve connectivity problems.
To test and verify your TCP/IP settings, you can use
the following utilities:
á IPCONFIG
á PING
The most effective method for troubleshooting connectivity is to understand thoroughly the installation and
configuration options of each of the network protocols.
If you understand the options available, you can narrow
down the possible problem areas very quickly. Also
ensure that you use utilities such as IPCONFIG and
PING to test your connections.
Choose the appropriate course of action to take to
resolve resource access and permission problems.
You should keep in mind two main issues about permissions:
á The default permissions for both share and
NTFS give the Windows NT group Everyone full
control over the files and folders. Whenever you
format a drive as NTFS or first share a folder, you
should remove these permissions. The Everyone
group contains everyone, including guests and
any other user who, for one reason or another,
can connect to your system.
á The NTFS folder permission delete takes precedence over any file permissions. In all other cases,
the file permissions take precedence over the folder permissions.
Choose the appropriate course of action to take to
resolve fault-tolerance failures. Fault-tolerance methods
include the following:
á Tape backup
á Mirroring
á Stripe set with parity
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 507
FAST FACTS: WINDOWS NT SERVER 4 ENTERPRISE EX AM
In using the NTBACKUP tool, the primary thing that
you need to do is to determine the frequency and type
of backup that you will do. There are three main types
of backups that you might want to perform:
á Full. This backs up all the files that you mark,
and marks the files as having been backed up.
This is the longest of the backups because it
transfers the most data.
á Differential. This backs up all the files that have
changed since the last backup. A differential
backup does not mark the files as being backed
up. As time passes since the last full backup, the
differentials become increasingly larger. However,
you need only reload the full backup and the
differential to return to the position of the last
backup.
á Incremental. This backs up any files that have
changed since the last backup, and then marks
them as having been backed up. If your system
crashes, you need to start by loading a full backup and then each incremental backup since that
full backup.
If you are mirroring the system partition, the disks and
partitions should be absolutely identical. Otherwise,
the MBR/DBR (master boot record/disk boot record)
that contains the driver information will not be correct.
Although ARC naming looks complicated, it is really
rather simple. The name is in four parts, of which you
use three. The syntax is as follows:
multi/scsi(#)disk(#)rdisk(#)partition(#)
The following list outlines the parts of the name:
á multi/scsi. You use either multi or scsi, not both.
Use multi in all cases except when using a scsi
controller that cannot handle int13 (hard disk
access) BIOS routines. Such cases are uncommon. The number is the logical number of the
controller with the first controller being 0, the
second being 1, and so forth.
507
á disk. When you use a scsi disk, you use the disk
parameter to indicate which of the drives on the
controller is the drive you are talking about.
Again, the numbers start at 0 for the first drive
and then increase for each subsequent drive.
á rdisk. Use this parameter for the other controllers
in the same way as you use the disk parameter for
scsi.
á partition. This is the partition on the disk that
you are pointing at. The first partition is 1, the
second is 2, and so forth. Remember that you can
have up to four primary partitions, or three primary and one extended. The extended partition
is always the last one, and the first logical drive in
the partition will have the partition’s number.
Other drives in the extended partition each continue to add one.
Breaking a mirror set. The boot floppy will get the
operating system up and running. You should immediately back up the mirrored copy of the mirror set. To
back up the drive, you must break your mirror set. To
do this, perform the tasks outlined in Step by Step FF.1.
STEP BY STEP
FF.1 Breaking the Mirror Set
1. Run the Disk Administrator.
2. From the Disk Administrator, click the
remaining fragment of the mirrored set.
3. Choose Fault Tolerance, Break Mirror set
from the menu.
At the end of these three steps, you should
notice that the mirror set has been broken,
and you can now back up the drive.
14 918-1 Fast Facts ent.i 8/28/98 1:29 PM Page 508
508
FA S T FAC TS : W I ND OW S NT S E RVER 4 ENTERPRISE EXAM
Regenerating a stripe set with parity. Fixing a stripe
set with parity is simple. Perform the tasks outlined in
Step by Step FF.2 to regenerate your stripe set with
parity.
STEP BY STEP
FF.2 Regenerating the Stripe Set
1. Physically replace the faulty disk drive.
2. Start the Disk Administrator.
3. Select the stripe set with parity that you need
to repair and then Ctrl+click the free space of
the drive you added to fix the stripe set.
4. Choose Fault Tolerant, Regenerate. Note that
this process can take some time, although the
process takes less time than restoring from
tape.
The drives regenerate all the required data
from the parity bits and the data bits, and
upon completion your stripe set with parity
is completely functional.
á Tape backups. In any system that you are using,
ensure that you have a good backup strategy. Any
component in your system can be faulty, and it is
your responsibility to have a recovery plan in case
of emergencies.
á Disk mirroring. If you are implementing disk
mirroring in your system, ensure that you have
created a fault-tolerant boot disk that you can use
in case of drive failure. By having this disk preconfigured and handy, you can break the mirror
set and replace the drive with very little downtime for your server.
á Stripe set with parity. This system automatically
regenerates data