Cabletron Systems 802 Specifications

Cabletron Systems
ETHERNET TECHNOLOGY GUIDE
Notice
Cabletron Systems reserves the right to make changes in specifications and other information
contained in this document without prior notice. The reader should in all cases consult Cabletron
Systems to determine whether any such changes have been made.
The hardware, firmware, or software described in this manual is subject to change without notice.
IN NO EVENT SHALL CABLETRON SYSTEMS BE LIABLE FOR ANY INCIDENTAL, INDIRECT,
SPECIAL, OR CONSEQUENTIAL DAMAGES WHATSOEVER (INCLUDING BUT NOT LIMITED
TO LOST PROFITS) ARISING OUT OF OR RELATED TO THIS MANUAL OR THE INFORMATION
CONTAINED IN IT, EVEN IF CABLETRON SYSTEMS HAS BEEN ADVISED OF, KNOWN, OR
SHOULD HAVE KNOWN, THE POSSIBILITY OF SUCH DAMAGES.
Copyright  1997 by Cabletron Systems, Inc. All rights reserved.
Printed in the United States of America.
Order Number: 9031913-01 April 1997
Cabletron Systems, Inc.
P.O. Box 5005
Rochester, NH 03866-5005
Cabletron Systems, SPECTRUM, BRIM, DNI, FNB, INA, Integrated Network Architecture,
LANVIEW, LANVIEW Secure, Multi Media Access Center, and MicroMMAC are registered
trademarks, and Bridge/Router Interface Modules, BRIM-A100, CXRMIM, Desktop Network
Interface, Distributed LAN Monitoring, Distributed Network Server, DLM, EFDMIM, EMM-E6,
EMME, EPIM, EPIM-3PS, EPIM-A, EPIM-C, EPIM-F1, EPIM-F2, EPIM-F3, EPIM-T, EPIM-T1,
EPIM-X, ESXMIM, ETSMIM, ETWMIM, FDCMIM-04, FDCMIM-08, FDMMIM, FDMMIM-04,
Flexible Network Bus, FOMIM, FORMIM, HubSTACK, IRBM, IRM, IRM-2, IRM-3, Media
Interface Module, MIM, MMAC, MMAC-3, MMAC-3FNB, MMAC-5, MMAC-5FNB, MMAC-8,
MMAC-8FNB, MMAC-M8FNB, MMAC-Plus, MRX, MRXI, MRXI-24, MultiChannel, NB20E,
NB25E, NB30, NB35, NBR-220/420/620, RMIM, SecureFast Switching, SecureFast Packet Switching,
SFPS, SPECTRUM Element Manager, SPECTRUM for Open Systems, TPMIM, TPMIM-22,
TPMIM-T1, TPRMIM, TPRMIM-36, TPT-T, TRBMIM, TRMM-2, and TRMMIM are trademarks of
Cabletron Systems, Inc.
AppleTalk, Apple, Macintosh, and TokenTalk are registered trademarks; and Apple Remote Access
and EtherTalk are trademarks of Apple Computer, Inc.
Ethernet, NS, Xerox Network Systems and XNS are trademarks of Xerox Corporation.
i
Notice
ii
Contents
CHAPTER 1
OVERVIEW
Purpose of This Manual .........................................................................................................1-1
Who Should Use This Manual................................................................................................1-1
Structure of This Manual ........................................................................................................1-2
CHAPTER 2
INTRODUCTION
Ethernet History .....................................................................................................................2-1
Ethernet Features ..................................................................................................................2-2
Media Access Method .....................................................................................................2-2
Bandwidth ........................................................................................................................2-2
Transmission Medium ......................................................................................................2-3
Frame Transmission.........................................................................................................2-4
Ethernet Topologies .........................................................................................................2-4
Bus Topology.............................................................................................................2-5
Ring Topology............................................................................................................2-6
Star Topology ............................................................................................................2-6
Hybrid Network Topology ..........................................................................................2-7
CHAPTER 3
ETHERNET LAN STANDARDS
The Open Systems Interconnect (OSI) Model .......................................................................3-1
Application of the OSI Model ...........................................................................................3-2
CHAPTER 4
ETHERNET DATA FRAMES
Manchester Encoding ............................................................................................................4-1
Ethernet Data Frames............................................................................................................4-3
Data Frame Size ..............................................................................................................4-4
Data Frame Types............................................................................................................4-5
Ethernet II Frame Type ....................................................................................................4-6
Ethernet “Raw” Frame Type .............................................................................................4-6
Ethernet 802.2 Frame Type .............................................................................................4-7
Ethernet SNAP Frame Type.............................................................................................4-9
Ethernet Addressing Schemes ......................................................................................4-10
Specific Addressing.................................................................................................4-10
Multicast Addressing ...............................................................................................4-11
Broadcast Addressing .............................................................................................4-11
iii
Contents
CHAPTER 5
ETHERNET MEDIA ACCESS METHOD
Clean Frame Transmission .................................................................................................... 5-1
Packet Involved in a Collision ................................................................................................ 5-2
Collision Detection on Point-to-Point Media .................................................................... 5-3
Out-Of-Window Collision................................................................................................. 5-3
CHAPTER 6
ETHERNET DEVICES
Ethernet Stations ................................................................................................................... 6-1
Ethernet Transceivers ............................................................................................................ 6-2
Multi-port Transceivers .................................................................................................... 6-3
Ethernet Repeaters ............................................................................................................... 6-3
Repeaters and Collisions ................................................................................................ 6-4
Auto Partition................................................................................................................... 6-4
Multi-port Repeaters ....................................................................................................... 6-5
Inter-Repeater Links (IRLs)............................................................................................. 6-5
Ethernet Bridges.................................................................................................................... 6-5
Routers .................................................................................................................................. 6-6
CHAPTER 7
ETHERNET NETWORK DESIGN
10BASE5 Ethernet Network Design ...................................................................................... 7-1
Single Segment 10BASE5 Ethernet Network ................................................................. 7-1
Transceiver Placement ............................................................................................. 7-2
Multi-port Transceivers.............................................................................................. 7-3
Multi-port Transceiver Rules ..................................................................................... 7-4
Grounding and Insulation ......................................................................................... 7-5
Multiple Segment 10BASE5 Ethernet Network............................................................... 7-6
Repeater Use ........................................................................................................... 7-6
Inter-Repeater Link (IRL) .......................................................................................... 7-7
10BASE2 Ethernet Network Design ...................................................................................... 7-8
Single Segment 10BASE2 Ethernet Network ................................................................. 7-8
Workstation Connections.......................................................................................... 7-9
Grounding and Insulation ......................................................................................... 7-9
Multiple Segment 10BASE2 Ethernet Network............................................................... 7-9
Grounding and Insulation ....................................................................................... 7-11
Fiber Optic Ethernet Network Design .................................................................................. 7-11
10BASE-T Twisted Pair Ethernet Network Design .............................................................. 7-12
iv
Contents
CHAPTER 8
PROPAGATION DELAY
Calculating the Delay .............................................................................................................8-1
Propagation Delay Example...................................................................................................8-2
CHAPTER 9
ETHERNET BRIDGE OPERATION
Filtering and Forwarding ........................................................................................................9-1
Spanning Tree Algorithm ........................................................................................................9-3
Configuration BPDU ........................................................................................................9-4
Topology Change BPDU..................................................................................................9-5
Spanning Tree Operation .......................................................................................................9-6
Data Loop Resolution ....................................................................................................9-12
Index
v
Contents
vi
Chapter 1
Overview
Purpose of This Manual
Welcome to the Cabletron Systems Ethernet Technology Guide. This guide
discusses the aspect of an Ethernet network known as the physical and
datalink layer. Although there may be some mention of specific
networking products and software, the primary focus is on the
understanding, design and implementation of a generic Ethernet Local
Area Network (LAN).
Throughout this document, references are made to both Ethernet and
IEEE 802.3 (CSMA/CD). The differences between Ethernet Version 2 and
802.3 are relatively minor. Because of these minor differences you will find
the text often refers to 802.3 as Ethernet. Throughout the industry, the
most popular nomenclature for an IEEE 802.3 CSMA/CD network is
Ethernet. Where there is a major difference between Ethernet and 802.3, it
is noted. This manual provides a basic overview of Ethernet and IEEE
standard 802.3 technology and is not meant to be a complete guide. The
objective of this manual is to provide Cabletron Systems customers with
information to understand why networks should be designed in a
particular way and why the rules need to be followed.
Who Should Use This Manual
This manual is intended for users of Cabletron Systems Ethernet products
and should be used as a supplement to Cabletron Systems Ethernet
Product User’s Manuals.
1-1
Structure of This Manual
Structure of This Manual
This manual is organized as follows:
Chapter 1, Overview - Outlines the purpose of this manual, who should
use it, and how it is structured.
Chapter 2, Introduction - Introduces Ethernet and gives a brief discussion
of the features and characteristics of the Ethernet technology.
Chapter 3, Ethernet LAN Standards - Discusses the Open Systems
Interconnect Model (OSI).
Chapter 4, Ethernet Media Access Method - Explains the different Ethernet
data frames and describes the Ethernet data encoding technique.
Chapter 5, Ethernet Media Access Method - Provides a basic explanation of
Carrier Sense Multiple Access with Collision Detection (CSMA/CD)
operation.
Chapter 6, Ethernet Devices - Describes the different devices used on an
Ethernet network, and provides examples of each.
Chapter 7, Ethernet Network Design - Discusses the design considerations
of a single segment 10BASE5 Ethernet network and builds it into a multisegment Ethernet network. The chapter then continues with the design of
a 10BASE2 Ethernet network, a Fiber Optic Ethernet network, and finally
a 10BASE-T Ethernet network.
Chapter 8, Propagation Delay - Provides an example of the step-by-step
process used in calculating the propagation delay of an Ethernet network.
Chapter 9, Ethernet Bridge Operation - Provides a step-by-step explanation
of the operation and learning process of an Ethernet Bridge. It also
provides a detailed explanation on the Spanning Tree Process.
1 -2
Overview
Chapter 2
Introduction
This chapter introduces Ethernet features and describes characteristics that distinguish Ethernet from
other Local Area Network (LAN) technologies such as Token Ring or FDDI.
Ethernet History
Ethernet was developed by Xerox Corporation’s Palo Alto Research
Center (PARC) in the mid-1970s. Ethernet was the technological basis for
the IEEE 802.3 specification, which was initially released in 1980. Shortly
thereafter, Digital Equipment Corporation, Intel Corporation, and Xerox
Corporation jointly developed and released an Ethernet specification
(Version 2.0) that is compatible with IEEE 802.3. Together, Ethernet and
IEEE 802.3 currently maintain the greatest market share of any Local Area
Network (LAN) protocol. Today, the term Ethernet is often used to refer to
all Carrier Sense Multiple Access/Collision Detection (CSMA/CD) LANs
that generally conform to Ethernet specifications, including IEEE 802.3.
At the time of its creation, Ethernet was designed to fill the middle ground
between long-distance, low-speed networks carrying data at high speeds
for very limited distances. Today, Ethernet is well-suited to applications
where a local communication medium must carry sporadic, occasionally
heavy traffic at high peak data rates.
2-1
Ethernet Features
Ethernet Features
The Institute of Electrical and Electronic Engineers (IEEE) is a standards
organization that establishes standards for many different technical areas.
This broad standards responsibility includes computer networking. IEEE
project 802 is responsible for networking standards for all network access
methods while project 802.3 specifically defines the Carrier Sense Multiple
Access/Collision Detection (CSMA/CD) access method, or Ethernet.
The following sections detail specific features involved with the
CSMA/CD media access method.
Media Access Method
As mentioned above, Ethernet is a Carrier Sense Multiple
Access/Collision Detection (CSMA/CD) LAN technology. Stations on an
Ethernet LAN can access the network at any time. Before sending data,
Ethernet stations “listen” to the network to see if it is already in use. If so,
the station wishing to transmit waits. If the network is not in use, the
station transmits. A collision occurs when two stations listen for network
traffic, “hear” none, then transmit simultaneously. In this case, both
transmissions are damaged and the stations, sensing this collision, must
retransmit at some later time. Backoff algorithms determine when the
colliding stations retransmit.
Ethernet is a broadcast network. In other words, all stations see all frames,
regardless of whether they represent an intended destination. Each station
must examine received frames to determine if it is the destination. If so,
the frame is passed to a higher protocol layer for appropriate processing.
Bandwidth
Ethernet bandwidth is 10 megabits per second although later
developments have produced Fast-Ethernet bandwidths of 100 megabits
per second.
2-2
Introduction
Ethernet Features
Transmission Medium
Ethernet transmits data frames over a physical medium of coaxial, fiber
optic, or twisted pair cable. The coaxial and fiber optic cable typically
represents the backbone of an Ethernet LAN while twisted pair is used as
a low cost connection from the backbone to the desktop.
Ethernet LANs have the following media restrictions in order to adhere to
IEEE standards:
•
Bus Length: The maximum bus length for an Ethernet LAN for all
media types is the following:
-
NOTE
NOTE
These media lengths are not precise values. Actual maximum cable
lengths are strongly dependent on the physical cable characteristics.
•
AUI Length: The maximum Attachment Unit Interface (AUI) cable
length is 50 m for connections from a transceiver to an Ethernet device
and 16.5 m for office AUI.
•
Number of Stations per Network: IEEE standards specify that the
maximum allowable number of stations per un-bridged network is
1,024, regardless of media type. 10BASE5 networks are allowed 100
taps per segment while 10BASE2 networks are allowed 30 taps per
segment with a maximum of 64 devices per tap each (Fiber optic and
twisted pair cable are point-to-point media which do not allow taps or
branches).
If it becomes necessary to extend the network beyond the IEEE limit of
1,024 devices, a bridge can be used to connect another full specification
Ethernet network.
•
Introduction
500 m for 10BASE5 coaxial cable
185 m for 10BASE2 coaxial cable
2,000 m for multimode fiber optic (10BASE-F) cable (5,000 m for
single mode)
100 m for twisted pair (10BASE-T) cable.
Maximum Signal Path: The maximum allowable signal path is 4
repeaters, 5 segments (with at least 2 segments being unpopulated
Inter-Repeater Links), and 7 bridges for all media types.
2-3
Ethernet Features
Frame Transmission
Ethernet stations encode their data into groups using a method known as
Manchester Encoding. These data groups, called frames, can be one of
four basic Ethernet frame types:
•
802.3 “raw” frame
•
Ethernet II (DIX) frame
•
Ethernet 802.2 frame
•
Ethernet SNAP frame
The Ethernet 802.2 and Ethernet SNAP frames are extensions of the 802.3
“raw” frame format, while the Ethernet II frame is formatted differently.
More about Ethernet frame types and Manchester Encoding will be
discussed in Chapter 4, Ethernet Data Frames.
Ethernet Topologies
The physical topology of the network defines its shape. Various network
topologies exist in the form of stars, trees, rings or buses. Complex
networks may employ several of the above topologies to form a hybrid
topology. A bus of stars or a ring of buses are two such examples.
Figure 2-1 shows some examples of several topology types. The boxes
indicate equipment of some type and the lines indicate cabling.
Bus
Star
Ring
Hybrid
Tree
1913-01
Figure 2-1. Various Network Topologies
2-4
Introduction
Ethernet Features
The most popular topologies used in Ethernet are the bus, star and tree.
Even though this discussion is about Ethernet, it is worth spending a few
moments on topologies not commonly used with Ethernet. A couple of
definitions at this point will assist with understanding the following
descriptions.
•
Point-to-Point: A point-to-point connection is a connection between
two and only two network devices (computers, servers, printers, etc.).
No taps or daisy chains are allowed. The most common point-to-point
medias are twisted pair and fiber optics.
•
Multi-point: A multi-point connection utilizes a single cable to connect
more than two network devices. A cable that has several devices
connected, one after another, (also known as daisy chaining) is an
example of a multi-point connection. The most common multi-point
media is coaxial cable.
Today’s networks employ various media topologies. The following
sections look at the general characteristics of the three most popular
topologies: bus, ring, and star.
Bus Topology
Employed in most older networks, the bus topology is a multi-point
network topology is which all devices are connected by a single common
cable or communications link. Taps are used to get the signal from the
coaxial cable to the device.
General Characteristics
Ethernet uses what is known as a contention bus topology. Any station on
the network can talk as long as no other station is talking. The more
stations that are on the network and want to talk, the worse the overall
performance. As a general rule, bus topologies are fairly straightforward
and easy to expand. Often, it is possible to expand the network without
affecting the network operation.
Vulnerability
Because all stations in the network share a common transmission media, a
failure of that media interrupts the exchange of signals between stations.
In a properly designed and constructed network however, such failures
are uncommon. Failure of a single workstation will not usually affect the
entire network.
Introduction
2-5
Ethernet Features
Ring Topology
A ring topology is a point-to-point topology in which the network devices
are connected, device to device, in an unbroken circle. Each signal to be
transmitted on the network must be processed by each station on the ring
before it is passed (or repeated) to the next station.
General Characteristics
Ring topologies commonly use an access method that is called token
passing. No station may talk unless that station has a free token, or
specialized signal code designated to determine which station on the
network is allowed to transmit. The token is passed from station to station
on the network along with the data being transmitted until it is released
by the receiving station.
Ring topologies can be complex in nature. They are easy to expand but
may involve calculations of cable lengths to keep the network within
specification. Most modern ring topologies resemble a physical star but
careful examination will reveal a logical ring cabled in a star
configuration. The use of networking hardware such as modular hubs
takes care of maintaining the logical ring in the wiring closet.
Vulnerability
Adding or removing network stations is simple and can, in most cases, be
done while the network is in operation. High level software takes care of
the recognition of problem nodes and also, in most cases, will remove the
problem nodes from the network and automatically reconfigure the ring.
Star Topology
A star topology is a point-to-point network in which the network devices
are connected through a central concentrator or controller. Two types of
access methods are employed: polling and contention.
Polling Star Topology
On a polling network, devices cannot talk or send messages unless they
are given permission (or polled) by a central computer or controller. A
device must wait to transmit until the controller asks for the information.
Performance of a polled network is dependent on the performance of the
controller and the number of devices attached to the controller.
Failure of the controller in a star topology network will bring the network
down. Failure of an individual node typically will not affect the remainder
of the network.
2-6
Introduction
Ethernet Features
Contention Star Topology
The contention star is the access method used with Ethernet. Workstations
are connected to a hub or concentrator located in a wiring closet.
Contention rules dictate that only one station can transmit data at any
given time and any station may talk providing the network is quiet. This
access method eliminates the need for polling and vastly improves
throughput and performance. Hubs can be expanded to handle hundreds
of devices without performance degradation. Expansion is easily
accomplished by simply plugging in a connection at the concentrator.
Failure of the hub can bring that section of the network down. Some
manufacturers allow for redundant backup of the hub and multiple load
sharing power supplies to reduce the possibility of hub failure and
minimize the impact of any such failure. The failure of a node will not
normally affect network operation.
Hybrid Network Topology
A hybrid topology is a combination of any of the three major topologies.
Examples include a ring of stars or a bus of stars or trees. Hybrid networks
may use a combination of point-to-point and multi-point connection
techniques.
For any of the above networks to function reliably and to allow multiple
vendors’ equipment to interoperate on the same network, a set of
standards have been developed. The most notable standards
organizations affecting data communications are the International
Organization for Standardization (ISO), the Institute of Electrical and
Electronic Engineers (IEEE), the Telecommunications Industry Association
(TIA), the American National Standards Institute (ANSI), and the
Consultative Committee on International Telegraphy and Telephony
(CCITT). In the following chapter we will look at the ISO standard as well
as the IEEE project 802 standards, specifically 802.3.
Introduction
2-7
Ethernet Features
2-8
Introduction
Chapter 3
Ethernet LAN Standards
Standards play an important role in modern local area networks. Without standards, users are forced
to buy proprietary networking equipment from a single vendor. Companies come and go, and product
lines are changed or discontinued. This leads to increased network costs to the users, network down
time and network equipment that does not inter-operate if standards are not in place.
Standards allow for easy integration of multiple vendor equipment into a common network. If one
company disappears, a purchaser has the flexibility to purchase another vendor’s product, being
confident that the two standards based products inter-operate.
The Open Systems Interconnect (OSI) Model
The International Organization for Standardization (ISO) Open Systems
Interconnect (OSI) Model provides a framework for the development of
system connection standards by using a consistent hierarchy of rules. The
OSI model defines where the needed tasks for system interconnection are
performed but not how they are performed. How tasks are performed on
a given layer is defined by the protocols, or rules, written for that
particular network based on the OSI model. The layers may be
implemented in hardware, software or both. Each layer in a network
based on the OSI Model performs specific functions required for proper
system interconnection.
As shown in Figure 3-1, there are seven layers in the OSI Model. They
begin with the Application Layer and finish with the Physical Layer.
3-1
The Open Systems Interconnect (OSI) Model
7
Application
6
Presentation
5
4
3
2
1
Session
Transport
Network
Data Link
Physical
1913-02
Figure 3-1. Open Systems Interconnect (OSI) Model
Application of the OSI Model
The perception of network operation appears as a direct peer-to-peer
communication to the user. The user message appears to go from the
sending application layer directly to the receiving application layer as if
the devices were directly attached. In actuality, the user message is routed
from the sending application layer down through the other layers of the
system. Each layer adds to or modifies the message according to the
network operating system’s protocol for each layer. The message passes
through all the layers of the system before appearing on the data channel
(cable or communications media) at the physical layer.
From the data channel the message passes upward through the same
layers at the destination device. As the message progresses from layer to
layer, each layer strips off information that was added by its counterpart
in the transmitting station. The result is the same message as was
originally sent, arriving at the top of the destination application layer.
Each layer performs a specific function with respect to the complete
communications process. The functions of each layer are as follows:
•
3-2
LAYER SEVEN: Application Layer–The application layer is the user’s
interface with the network. This layer directly interacts with user
application programs to provide access to the network. All other layers
exist to support the requirements of this layer. The application layer is
usually involved with tasks such as electronic mail and file transfer.
Ethernet LAN Standards
The Open Systems Interconnect (OSI) Model
•
LAYER SIX: Presentation Layer–The presentation layer deals with data
translation and code conversion between devices with different data
formats (e.g., ASCII to EBCDIC). This layer also handles translation
between differing device types and file formats, as well as data
encryption and decrypting services.
•
LAYER FIVE: Session Layer–The session layer manages the
communication dialogue (the “session”) between two communicating
devices. The session layer establishes rules for initiating and
terminating communications between devices and provides error
recovery as well. If an error or communications failure is detected, the
session layer retransmits data to complete the communications process.
The session layer requests a certain level of service from the transport
layer such as one way transmission that doesn’t require a reply, or a two
way conversation that requires a lot of monitoring and feedback.
•
LAYER FOUR: Transport Layer–The transport layer deals with the
optimization of data transfer from source to destination by managing
network data flow and implementing the quality of service requested
by the session layer. The transport layer determines the packet size
requirements based on the amount of data to be sent and the maximum
packet size allowed on the communications media. If the data to be sent
is larger than the maximum packet size allowed on the network, the
transport layer is responsible for dividing the data into acceptable sizes
and sequencing each packet for transmission. During the dividing and
sequencing process, this layer adds information such as sequence
number and error control information to the data portion of the packet.
When receiving data from the network layer, the transport layer
ensures that the data is received in order and checks for duplicate and
lost frames. If data is received out of order, which is possible in a
larger, routed network, the transport layer correctly orders the data
and passes the data up to the session layer for additional processing. A
popular protocol that uses the transport layer is Transmission Control
Protocol (TCP) used in TCP/IP.
•
LAYER THREE: Network Layer–The network layer accepts data from
the transport layer and adds the appropriate information to the packet
to provide proper network routing and some level of error control. Data
is formatted for the appropriate communications method such as local
area network, wide area network such as T1, or packet switched
technology such as X.25. A popular protocol that uses the network layer
is the Internet Protocol (IP) used by TCP/IP.
Ethernet LAN Standards
3-3
The Open Systems Interconnect (OSI) Model
•
LAYER TWO: Data Link Layer–The data link layer is involved with
transmission, error detection and flow control of the data. The major
function of the data link layer is to act as a shield for the higher layers
of the network model, controlling the actual transmission and
reception process. Error detection and control of the physical layer are
the primary functions of this layer ensuring the upper layers that any
data received from the network is error free. The IEEE 802 model
divides the data link layer into two sub-layers: Logical Link Control
(LLC) and Media Access Control (MAC).
6. Presentation
7
6
5. Session
5
4. Transport
4
3
LLC
2
1
MAC
7. Application
3. Network
2. Data Link
1. Physical
1913-03
Figure 3-2. Data Link Layer of the OSI Model
3-4
-
Logical Link Control–The LLC layer is responsible for shielding
the upper layers from any particular access method or media. The
upper layers need not worry about whether they are connected to a
Token Ring or Ethernet network because the logical link control
handles the interface. The LLC is a defined standard (IEEE 802.2)
that provides for a common interface of the layers above to any
physical network implementation.
-
Media Access Control–The MAC layer is responsible for several
areas of operation. On the transmit side the MAC layer is
responsible for receiving data from the Logical Link Control layer
and encapsulating it into a packet ready for transmission. The MAC
layer is also responsible for determining if the communications
channel is available. If the channel is available, the MAC layer
transmits the data onto the cable through the physical layer and
monitors the physical layer status for an indication of a collision
(more than one station transmitting at the same time). If there is a
collision, the MAC layer also handles the backoff and
retransmission function.
Ethernet LAN Standards
The Open Systems Interconnect (OSI) Model
On the receive side, the MAC layer follows the reverse of the above
steps. It checks the frame for errors, strips control information then
passes the remainder of the packet to the upper layers by way of
logical link control.
•
LAYER ONE: Physical Layer–At this layer, the transmission of data
between devices is defined. The definition includes cables and
connectors, connector pinouts, voltage levels that represent digital
logic levels, bit timing, and the actual network interface device called a
Transceiver (transmitter/receiver). The IEEE 802 model divides the
physical layer into four sub-layers: Physical Layer Signaling (PLS),
Attachment Unit Interface (AUI), Physical Medium Attachment
(PMA), and Medium Dependent Interface (MDI).
6. Presentation
7
6
5. Session
5
4. Transport
4
3
7. Application
3. Network
2. Data Link
1. Physical
2
1
PLS
AUI
PMA
MDI
1913-04
Figure 3-3. Physical Layer of the OSI Model
-
PLS (Physical Layer Signaling)–Defines the signaling and the
interface to the transceiver cable.
-
AUI (Attachment Unit Interface)–Defines the transceiver cable
specifications.
-
PMA (Physical Medium Attachment)–Defines the transceiver
operation and specifications.
-
MDI (Medium Dependent Interface)–Defines the specifications
for the portion of the transceiver that connects to specific cable
types such as 10BASE5 coaxial cable.
The remaining chapters will concentrate primarily on the Physical and
Data Link layers of the OSI Model.
Ethernet LAN Standards
3-5
The Open Systems Interconnect (OSI) Model
3-6
Ethernet LAN Standards
Chapter 4
Ethernet Data Frames
Most computer networks require that information transmitted between two stations be divided into
blocks called frames. For these frames to be sent successfully to other devices on the network, certain
protocol and routing information must be added to the data. In addition, the way this information is
arranged inside the frame must conform to a specific format. The way an Ethernet device places the
data bits into frames before it is placed on the LAN is called Manchester Encoding.
The following chapter discusses Manchester Encoding and describes the four different Ethernet frame
types.
Manchester Encoding
The information that is to be transmitted on the cable is in the form of a
constantly changing voltage signal. This signal is electronically
transformed into a D. C. signal with a value of either 0 volts or -1.2 volts.
The value of the D. C. signal is found by periodically sampling the voltage
value of the original signal and assigning it one of the two values,
depending on the value found at the time it was checked. The result is to
change a non-constant analog signal into a digital signal with a value
toggling between either 0 or -1.2 volts.
4-1
Manchester Encoding
In Ethernet, the digital D.C. signal is transformed into discrete time
segments called bits by a method called Manchester Encoding. With
Manchester Encoding, the incoming digital signal is checked at specific
time intervals for its change of state. In other words, the signal is checked
to see if it is changing from 0 volts to -1.2 volts or from -1.2 volts to 0 volts
during a certain time period. Depending on its change of state in this
specific time interval, or bit time, the signal is assigned a logic “1” or a
logic “0” for that time interval. The resulting signal is a steady stream of
bit times (or bits) with values of either logic “1” corresponding to a change
from -1.2 v to 0 v or a logic “0” corresponding to a change from 0 v to
-1.2 v. If there is no change of state during a certain bit time, that bit time is
assigned a value corresponding to the value of the bit time preceding it.
Manchester Encoding also provides the digital signal with a method of
alignment by ensuring that the transitions to the logic levels happen only
during the center of the bit time. Figure 4-1 shows how Manchester
Encoding works on a random digital signal.
Changes to logic levels take place
at the center of each bit time.
0
0
1
1
0
0 volts D.C.
-1.2 volts D.C.
Bit Time
.1 µs
no change: value
stays the same as
preceding bit
Logic "0"
Logic "1"
1913-05
Figure 4-1. Manchester Encoding
4-2
Ethernet Data Frames
Ethernet Data Frames
Ethernet Data Frames
Previously it was shown how Ethernet data signals are transformed into
bits. Before these bits are sent onto an Ethernet network, they must be
formatted into specific groups called data frames. Data frames are strings
of bytes (eight bits equal one byte) which contain addressing, timing,
protocol, and error correction information as well as the data being sent.
The packet structure used in IEEE 802.3 and Ethernet is shown in
Figure 4-2.
Data Frame
Preamble SFD
Dest.
Address
Source
Address
Length
Data
7 Bytes 1 Byte
6 Bytes
6 Bytes
2 Bytes
46-1500 Bytes
CRC
4 Bytes
1913-06
Figure 4-2. Ethernet Data Frames
Each section of the frame is described as follows:
•
Preamble: The preamble indicates the beginning of frame transmission.
The preamble allows for frame timing at the receiving station. The
signal pattern is a repeating pattern of alternating ones and zeroes for
a total of 56 bits (7 bytes).
•
Start Frame Delimiter (SFD): The SFD signal pattern is 10101011 for a
total of 8 bits (1 byte). It follows the preamble and indicates the start of
information by the last two bits, 11.
•
Destination Address: The address of the station, or stations, that the
data frame is intended for. It follows the SFD and is 48 bits (6 bytes) in
length.
•
Source Address: Follows the destination address and indicates the
address of the station initiating the transmission. The source address is
48 bits (6 bytes) in length.
•
Length Field: The length field follows the source address and indicates
the length of the data field. The length field is 16 bits (2 bytes) long. In
Ethernet version 1.0 or version 2.0, this field is called a type field. The
type field will usually indicate the packet protocol (e.g., TCP/IP, XNS,
DECNet, Novell IPX, etc.).
Ethernet Data Frames
4-3
Ethernet Data Frames
•
Data Field: The data field follows the length field. It is 46 bytes
minimum to a maximum of 1500 bytes in length. This field contains the
actual data being sent across the network along with some control
information. If the data to be sent is less than the minimum 46-byte
packet size, a special bit pattern called PAD is used to fill in up to the
46-byte minimum. The minimum packet size set by the IEEE 802.3
specification is explained below.
•
Cyclic Redundancy Check (CRC): The CRC follows the data field and
is 32 bits (4 bytes) in length. Also knows as the Frame Check Sequence
(FCS), this field is used to check the integrity of the frame. Before
placing a frame out on the wire, the sending station takes all the bytes
within the frame, performs a mathematical calculation, and places the
result at the end of the frame. When the frame arrives at the destination,
the receiving station performs the same mathematical calculation and
should receive the same result. If not, it assumes something has been
corrupted and discards the frame.
Data Frame Size
IEEE defines both a minimum and a maximum frame size. The minimum
frame size is 64 bytes (12 address bytes, 2 length bytes, 46 data bytes and 4
CRC bytes). The maximum frame size is 1,518 bytes (same as above with
1,500 byte data field).
The minimum frame size has been determined to give the best bridge
switch speed on heavily used networks by minimizing the amount of time
a station must defer to other transmissions. It also increases the amount of
overhead involved in completing a transmission. The minimum frame
size will move quickly because of its size, but two full size frames move as
much information as 66 minimum size frames containing only 46 bytes of
data each. Therefore, the two large frames require only 36 bytes of
overhead (Preamble, SFD, addresses, etc.) while the small frames require
1,188 bytes of overhead: nearly half the size of the original transmission.
This also doesn’t include the problems involved in trying to transmit 66
times in an operating network with collisions. The minimum frame size
also plays an important role in the detection of collisions and determining
the maximum network size.
In an Ethernet network, a station must still be transmitting its data to
detect that it was involved in a collision. We know that a minimum size
frame is 64 bytes in length which equates to 512 bits (8 bits per byte). We
also know that each bit time in an Ethernet network is defined as 0.1 µs.
Multiplying 512 bits by 0.1 µs yields 51.2 µs to transmit a 64-byte
minimum size frame.
4-4
Ethernet Data Frames
Ethernet Data Frames
Due to inherent propagation delays in electronics and cabling it would
make sense that within 25.6 µs (half of 51.2 µs) our transmitted signal
should have reached the farthest point on the network. If a collision were
to happen at the farthest point on the network the collision signal will
have the remaining 25.6 µs to travel back to the transmitting node thus
alerting the node that its transmission needs to be re-sent. The 25.6 µs one
way propagation window is also called the collision domain.
Data Frame Types
Ethernet data frames are packaged one of four ways:
•
Ethernet II (DIX)
•
802.3 “raw”
•
Ethernet 802.2
•
Ethernet SNAP
Figure 4-3 shows each of the Ethernet frame types. The Ethernet 802.2 and
Ethernet SNAP frames are extensions of the 802.3 “raw” frame format,
while the Ethernet II frame is formatted differently. The following sections
describe each frame type.
Ethernet II Frame
Preamble
Destination
Address
Source
Address
Type
Data
FCS
8 bytes
6 bytes
6 bytes
2 bytes
46 - 1,500 bytes
4 bytes
802.3 "Raw" Frame
Preamble
Start Frame
Delimiter
Destination
Address
Source
Address
Length
Data
FCS
7 bytes
1 byte
6 bytes
6 bytes
2 bytes
46 - 1,500 bytes
4 bytes
Ethernet 802.2 Frame
Preamble
Start Frame
Delimiter
Destination
Address
Source
Address
Length
DSAP
SSAP
Control
Data
FCS
7 bytes
1 byte
6 bytes
6 bytes
2 bytes
1 byte
1 byte
1 byte
43 - 1,497 bytes
4 bytes
Ethernet SNAP Frame
Preamble
Start Frame
Delimiter
Destination
Address
Source
Address
Length
DSAP
SSAP
Control Protocol
Identifier
7 bytes
1 byte
6 bytes
6 bytes
2 bytes
1 byte
1 byte
1 byte
5 bytes
Data
FCS
37 - 1,492 bytes
4 bytes
1913-07
Figure 4-3. Ethernet Frame Types
Ethernet Data Frames
4-5
Ethernet Data Frames
Ethernet II Frame Type
In the early days of computer networks, Digital, Intel, and Xerox got
together and specified a networking standard that they called Ethernet.
This standard included the definition of a data link level access method
and a packet format that shared the Ethernet name (it is now called
Ethernet II because it is in its second revision). Table 4-1 shows the fields
defined for Ethernet II data frames.
Table 4-1. Ethernet II Frame Type
Field Name
Field Size
Field Definition
Preamble
8 bytes
Signals beginning of the packet.
Destination
Address
6 bytes
Contains address of the destination of the
frame.
Source Address
6 bytes
Contains address of the packet’s origin.
Type Field
2 bytes
Specifies the upper layer protocol used.
Data
46–1500 bytes
Contains the data to be transferred.
Frame Check
Sequence
4 bytes
Verifies the integrity of the frame.
The Ethernet II standard specified that a header (consisting of the
Preamble, Source and Destination addresses, and frame type) be added to
the data before sending it on the network medium. The frame format
follows the rules to access a network using the CSMA/CD access method.
Ethernet “Raw” Frame Type
Eventually, both the Ethernet media and packet format were pursued by
the standards committees of the IEEE. Working from the original DIX
specification, IEEE proposed its own Ethernet standard which they called
802.3 (named after the committee that worked on it). Table 4-2 describes
each of the Ethernet 802.3 frame fields (also known as “raw” frames).
4-6
Ethernet Data Frames
Ethernet Data Frames
Table 4-2. Ethernet “Raw” Frame Type
Field Name
Field Size
Field Definition
Preamble
7 bytes
Signals beginning of the frame.
Start Frame
Delimiter
1 byte
Signals start of data.
Destination
Address
6 bytes
Contains the address of the destination of
the frame.
Source Address
6 bytes
Contains the address of the frame’s
origin.
Length Field
2 bytes
Specifies the length of the data field.
Data
46–1500 bytes
Contains the data to be transferred.
Frame Check
Sequence
4 bytes
Determines the integrity of the frame.
The IEEE 802.3 frame format is almost identical to the Ethernet II format.
The only difference is that a length field is used in place of the type field.
This field indicates the length of the data portion of the 802.3 frame, with
the maximum length being 1,518 decimal.
A network device can decipher the difference between an 802.3 “raw”
frame and an Ethernet II frame by looking at this portion of the packet (the
length or type field). As it turns out, the assigned values for the Ethernet II
type field are always greater than 1,500 decimal. Since the maximum
frame size for Ethernet is 1,518 bytes, with 18 bytes of overhead, the length
field always contains a value less than that.
Ethernet 802.2 Frame Type
Without a protocol type field, it is impossible to determine what protocol
to use for interpreting the encapsulated data in an 802.3 “raw” frame. If
more than one upper-layer protocol exists on the network, the packet may
be incorrectly routed. Therefore, sometime after the 802.3 standard was
released, IEEE came out with the 802.2 standard. Table 4-3 describes each
of the Ethernet 802.2 frame fields.
Ethernet Data Frames
4-7
Ethernet Data Frames
Table 4-3. Ethernet 802.2 Frame Type
Field Name
Field Size
Field Definition
Preamble
7 bytes
Signals beginning of the frame.
Start Frame
Delimiter
1 byte
Signals start of data.
Destination
Address
6 bytes
Contains address of the destination
frame.
Source Address
6 bytes
Contains address of the frame’s origin.
Length Field
2 bytes
Indicates length of the Data plus LLC
fields.
Destination
Service Access
Point (DSAP)
1 byte
Shows first byte of 2 byte value indicating
the frame’s upper layer protocol
destination.
(Source Service
Access Point) SSAP
1 byte
Shows second byte of 2 byte value
indicating the frame’s upper layer
protocol destination.
Control
1 byte
Indicates the type of LLC frame.
Data
43–1,497
bytes
Contains the data to be transferred.
Frame Check
Sequence
4 bytes
Determines the integrity of the frame.
The Ethernet 802.2 header envelopes the data before it is encapsulated
within an IEEE 802.3 header. This frame adds several fields to the header;
a destination service access point (DSAP), a source service access point
(SSAP), and a control field.
4-8
•
Service Access Point: This denotes the point of service the packet is
intended for, or what upper layer protocol is supposed to use the data.
Both the DASP and the SSAP fields contain values that identify the
upper layer protocol type of the frame.
•
Control Field: This is used by certain protocols for administrative
purposes.
Ethernet Data Frames
Ethernet Data Frames
Ethernet SNAP Frame Type
After the 802.2 frame was defined, there was some concern that the one
byte DSAP and SSAP fields were not adequate for the number of protocols
that eventually needed to be identified. In response from Apple Computer
and the TCP/IP community, another frame standard was defined for both
Ethernet and Token Ring. It was called the Sub-Network Access Protocol,
shown in Table 4-4 below.
Table 4-4. Ethernet SNAP Frame Type
Field Name
Field Size
Field Definition
Preamble
7 bytes
Signals beginning of the frame.
Start Frame
Delimiter
1 byte
Signals start of data.
Destination
Address
6 bytes
Contains the address of the destination of
the frame.
Source Address
6 bytes
Contains the address of the frame’s
origin.
Length Field
2 bytes
Contains the length of the Data plus LLC
fields.
Destination
Service Access
Point (DSAP)
1 byte
Set to AA (hex) and 10101010 (binary).
Source Service
Access Point
(SSAP)
1 byte
Set to AA (hex) and 10101010 (binary).
Control Field
1 byte
Set to 03 (hex) and 00000011 (binary).
Protocol Identifier
5 bytes
Specifies the upper layer protocol.
Data
38– 1,492
bytes
Contains the data to be transferred.
Frame Check
Sequence
4 bytes
Determines integrity of the frame.
Ethernet Data Frames
4-9
Ethernet Data Frames
This frame type adds a five-byte protocol identification field at the end of
the 802.2 header, where the protocol is identified. To distinguish an IEEE
802.2 SNAP frame, the value of the DSAP and SSAP fields in the 802.2
header are both set to AA. If a network device finds AA in the DSAP and
SSAP fields, it knows this is a SNAP-based frame and it should look for
the protocol type in the protocol identification field.
Ethernet Addressing Schemes
There are three types of addressing schemes used in Ethernet networks.
Each address type serves a different purpose. They are as follows:
1. Specific Addressing
2. Multicast Addressing
3. Broadcast Addressing
Specific Addressing
The IEEE specifies that each addressable network device will have a
unique hardware address that is made up of 6 bytes of information. The
address is either hard-coded into every network interface controller card
during manufacturing or assigned out of a group of addresses given to a
particular corporation. The availability of addresses is strictly controlled
by the IEEE.
The IEEE assigns each network hardware manufacturer a unique
manufacturer identifier and a block of numbers that the manufacturer
usually assigns sequentially to each piece of hardware. The combination
of the manufacturer ID and the sequential number makes up the common
48-bit Ethernet address.
The first 3 bytes of the address contains the manufacturer identifier and
the last 3 bytes contain the sequential numbering. The numbering scheme
is given in hexadecimal format. An example of a typical Ethernet address
is shown in Figure 4-4.
00-00-1D-00-26-A3
Manufacturer ID
Sequential Address
1913-08
Figure 4-4. A Typical Ethernet Address
4-10
Ethernet Data Frames
Ethernet Data Frames
The 00-00-1D manufacturer ID belongs to Cabletron Systems. When a
specific Ethernet address is used as the destination address in a packet,
that packet will be decoded only by the station with that specific address.
Multicast Addressing
At times it is necessary to communicate with many devices on a network
simultaneously. For instance, a network management station might poll or
query a group of devices to determine their status. Instead of keeping a list
or table of unique addresses a group address can be used. This type of
group addressing is accomplished using a multicast address.
A multicast address is formed by modifying the manufacturer ID portion
of the destination address. As shown in Figure 4-5, the least significant bit
of the first byte is changed from a 0 to a 1. The net result is to turn a
Cabletron Systems address of 00-00-1D... into a multicast address of
01-00-1D... A multicast address is destined for devices from a particular
manufacturer.
Address
00-00-1D-00-26-A3
Multicast Address
01-00-1D-00-26-A3
0000 0000
0000 0001
Lowest bit of the first byte
changes from 0 to 1
1913-09
Figure 4-5. Ethernet Multicast Address
Broadcast Addressing
A broadcast address is an address that is meant to be heard by all stations
on the network. Certain protocols running on workstations will
occasionally send out broadcast messages to servers on the network to let
the servers know that the node is on-line.
Ethernet Data Frames
4-11
Ethernet Data Frames
A broadcast address contains all “F” hexadecimal characters which is
equivalent to all bits being set to logic 1 in both the manufacturer ID and
sequential number area of the address (see Figure 4-6).
FF-FF-FF-FF-FF-FF
Manufacturer ID
Sequential Address
1913-10
Figure 4-6. Ethernet Broadcast Address
In the following chapter we will look at how the packet is transmitted onto
the network and the rules that must be followed to ensure a successful
transmission.
4-12
Ethernet Data Frames
Chapter 5
Ethernet Media Access Method
Ethernet, as stated in Chapter 1, uses a method of access control known as Carrier Sense Multiple
Access with Collision Detection, or CSMA/CD. Access to the network media is controlled by the lower
half of the Data Link Layer called Media Access Control, or MAC. The following chapter describes the
operation of the Ethernet MAC.
Clean Frame Transmission
The following determines the process that an Ethernet station goes
through in a clean frame transmission over the network.
A station wanting to transmit first listens to the communication channel to
see if any other station is transmitting. If a carrier is sensed (another
station is transmitting), the station waits a random length of time, and
then listens to the communication channel again. If no other station is
transmitting, the station begins frame transmission.
During frame transmission, the station continuously monitors the bus. As
long as the monitoring shows there is no other station transmitting on the
cable, the frame transmission continues until the transmission is complete.
Once the frame transmission is complete, the station is quiet for 9.6 µs to
allow for the required interframe gap. After 0.6 µs into the interframe gap,
the transceiver is given a 1.4 µs window to test its collision detect circuitry
(see Figure 5-1).
5-1
Packet Involved in a Collision
Interframe Gap
Data Packet
9.6 µs
.6 µs
1.4 µs
2 µs
SQE Test
1913-11
Figure 5-1. Ethernet Interframe Spacing
During this time, the station will see the Signal Quality Error (SQE) test
signal on its collision detection. When the station sees this signal during
the 1.4 µs window, it is informed that the transceiver collision detect
circuits are working properly and if a collision occurred, it would be
notified.
After the remainder of the 9.6 µs interframe gap is complete, the station is
able to start the transmission process again by first listening to see if the
cable is available.
Signal Quality Error Test
The SQE (Heartbeat) is only generated by a transceiver and is only seen by
the host device that is connected to that specific transceiver. The SQE does
not appear on the network bus. It is a signal from a transceiver to its
associated station simply to inform the station that the transceiver’s
collision detection is working properly.
Packet Involved in a Collision
Because of propagation delays in the network, it is possible for two
stations to simultaneously find the bus available, in which case both
stations will begin transmitting frames. When these signals meet on the
cable, a collision occurs. These two signal voltages add together and
increase the voltage level on the cable, which is sensed by the transmitting
transceivers. The transceiver then sends a collision signal to the host
station, all while it is still transmitting the packet. If the station is not still
transmitting the packet, there is a problem in the design of the network,
and it does not meet IEEE 802.3 specifications.
5-2
Ethernet Media Access Method
Packet Involved in a Collision
When a collision is detected, both stations will transmit a jam signal that is
long enough to ensure that the collision is detected by all stations on the
network. Then, each station involved in the collision will wait for a
random period of time and then attempt transmission again. The station
will attempt again transmission up to 16 consecutive times before an error
is sent to the upper layer protocols notifying the station of a serious
communication problem.
Collision Detection on Point-to-Point Media
On Point-to-Point media, it is not possible to detect a collision by listening
to the transmission as with multi-point media. Point-to-point media
transceivers use a method called Transmit Mode Collision Detection. With
this method, the transceivers will monitor their receive ports while
transmitting. If they receive a signal while transmitting, then a collision
has occurred.
Out-Of-Window Collision
As mentioned in Chapter 3, from any station on the network, a
transmitted frame has 25.6 µs to get to the end of the collision domain. If a
collision were to happen at the farthest point from the transmitting station
(25.6 µs away), the collision signal will take an additional 25.6 µs to
propagate back to the transmitting station for a total of 51.2 µs round trip
time (or the time it takes to transmit a minimum size frame). If a station is
able to transmit for 51.2 µs without detecting a collision, the station should
have acquired the communications channel and its signal should be the
only one using the network. If a collision is detected after the station has
transmitted the required minimum frame size, an Out-Of-Window (OOW)
collision has occurred. In other words, the station has transmitted for
51.2 µs without a collision but senses a collision after 51.2µs has passed.
Out- Of-Window collisions indicate abnormal network operation. They
are usually caused by the network being too long where the round trip
propagation delay is greater than 51.2 µs, a station somewhere on the
network is transmitting at will, or a cable somewhere on the network
failed during the transmission of the frame.
In the following chapters we will look at the importance of propagation
behavior. Propagation delay is discussed in more detail in Chapter 8,
Propagation Delay.
Ethernet Media Access Method
5-3
Packet Involved in a Collision
5-4
Ethernet Media Access Method
Chapter 6
Ethernet Devices
This chapter describes devices that are common to an Ethernet network. All devices attached to an
Ethernet bus must comply with the IEEE 802.3 Standard. Typical Ethernet devices include stations,
transceivers, repeaters, bridges and routers. Figure 6-1 shows various devices attached to an Ethernet
LAN. The following sections provide a description of each of these devices and their network functions.
Bridge
Network Bus
Network Bus
Repeater
Transceivers
AUI Cable
Stations
1913-12
Figure 6-1. Devices on an Ethernet LAN
Ethernet Stations
Ethernet stations are addressable nodes on an Ethernet network capable of
transmitting, receiving, and repeating information. Workstations, file
servers, and printers (shown in Figure 6-2) are some examples of these
types of devices. Stations connect to the Ethernet bus through devices
called transceivers and are discussed in the next section.
6-1
Ethernet Transceivers
Network
Server
Network
Printer
Station
1913-13
Figure 6-2. Ethernet Stations
Ethernet Transceivers
A transceiver (transmitter/receiver) is the device that connects
workstations, servers, and other equipment to the Ethernet cabling media
being used for network transmissions. For full descriptions of Ethernet
networking media, refer to the Cabletron Systems Cabling Guide. The
transceiver (Figure 6-3) is responsible for listening to the Ethernet bus to
determine if the bus is currently in use by another station. If a collision
occurs during transmission, the transceiver is responsible for alerting its
connected device by sending a collision signal down the AUI cable.
IEEE 802.3 compliant transceivers employ special circuitry to perform a
watchdog function on the transceiver transmitter. If a workstation starts to
continuously transmit data, the transmitter shuts down so the network is
not taken over by one device.
ER
SCEIV
TRAN
AUI Connector
1913-14
Figure 6-3. Standard Ethernet Transceiver
6-2
Ethernet Devices
Ethernet Repeaters
Multi-port Transceivers
A multi-port transceiver or fanout is a transceiver that has one port to
connect to a regular transceiver and up to fifteen AUI ports to connect
through AUI cables to individual devices. This allows you to connect
several addressable devices to one cable tap. IEEE 802.3 standards for
transceiver placement and tap spacing specify that only 100 taps will be
allowed on a 10BASE5 segment with a distance of 2.5 m between them. If
it becomes necessary to concentrate a number of workstations in one
physical location, a multi-port transceiver can be used.
Figure 6-4 shows a typical multi-port transceiver.
AC
1
2
3
4
5
6
7
8
Network
1913-15
Figure 6-4. Standard Ethernet Multi-port Transceiver
Ethernet Repeaters
If it is necessary to add additional taps beyond the 100 tap limit, another
coaxial segment must be added. If the new segment uses the same
architecture at the Physical Layer of the OSI model as the old segment, a
repeater can be used to join the two segments. A repeater regenerates the
preamble, and amplifies and retimes the signal from one cable segment to
the other.
Ethernet Devices
6-3
Ethernet Repeaters
Figure 6-5 shows a Cabletron Systems LR2000 two-port local Ethernet
repeater.
LR-2000
1913-16
Figure 6-5. Cabletron Systems Two-Port Local Ethernet Repeater
Repeaters and Collisions
A collision happens when more than one station transmits on the network
at one time. Since the repeater physically separates the two coaxial
segments, a collision on one segment cannot be seen by devices on the
other segment connected to the repeater. Therefore, the repeater is
responsible for ensuring that the collision signal is propagated to all
segments attached to it. To force a collision, the repeater sends out a
special bit pattern called a Jam signal to all segments attached to it. This
signal notifies all stations on the network that a collision has occurred.
Auto Partition
When the repeater detects 32 consecutive collisions on one port it will
logically turn off or segment the port that it detected the problem on, thus
allowing the rest of the network to function properly. When the repeater
detects a collision on the segmented port, the collision will not be
forwarded to the other segments of the network, leaving the port in
segmented condition. When the repeater receives a packet on a good port,
it attempts to transmit the packet to the segmented port. If the packet
transmits successfully, the repeater will turn the segmented port back on,
bringing it out of segmentation.
6-4
Ethernet Devices
Ethernet Bridges
Multi-port Repeaters
A multi-port repeater is a device which has more than two ports that
connect to full-length Ethernet Segments. Generally the repeaters are very
similar in appearance to the two-port local repeater shown in Figure 6-5,
with the exception of the number of ports. These repeaters regenerate the
preamble and amplify and re-time a signal from one cable segment to the
others.
Inter-Repeater Links (IRLs)
Up to 3 repeaters can be used to connect separate coaxial segments before
special considerations need to be taken into account. If the segments are
physically located a distance from each other, the use of an Inter-Repeater
Link (IRL) comes into play. As shown in Figure 6-6, an IRL is a segment
that connects only two repeaters. It can be made up of thin coaxial cable
(185 m), thick coaxial cable (500 m), twisted pair cable (200 m), or Fiber
optic cable (2/5 km). With the use of IRLs it is possible to build a network
consisting of up to three populated segments with two IRLs and four
repeaters joining them.
NETWORK 1
NETWORK 2
(Up to 1,024 devices)
(Up to 1,024 devices)
Inter Repeater Link
(length dependent
on media type)
Repeaters
1913-17
Figure 6-6. Using an IRL
Ethernet Bridges
A bridge is a device that can be added to a network to allow expansion
beyond the limitations of 802.3. If a network has a repeater hop of four
repeaters or a propagation delay near the 51.2 µs maximum, a bridge can
be used to accomplish the addition of the new full specification Ethernet.
Ethernet Devices
6-5
Routers
Unlike a repeater, which sends all frames it receives to all segments it is
connected to, a bridge reads the frames it receives and decides whether to
filter or forward the frame based on the addressing information contained
within it. Bridges can also be used to connect similar networks (networks
with the same upper five layers of the OSI model) such as Ethernet, Token
Ring, and Fiber Distributed Data Interface (FDDI) together. Because
bridges work at layer 2 of the OSI Model, they are protocol independent.
They have a longer response time than repeaters because a bridge must
read the complete data frame, check for errors, and make forward or filter
decisions based on recognized addresses stored in its source address table.
Bridges are discussed further in Chapter 9, Ethernet Bridge Operation.
Routers
A router works much like a bridge, except that a router pays attention to
the upper network layer protocols (OSI model level 3) rather than just
Physical Layer protocols like a bridge. A router will decide whether to
forward a packet by looking at the protocol level addresses rather than the
MAC address. Because routers transfer packets between different media
types, many routers can also function as bridges.
A detailed discussion of routing operations is beyond the scope of this
guide.
The following chapter explains the use of each of these devices while
considering your Ethernet network design.
6-6
Ethernet Devices
Chapter 7
Ethernet Network Design
Designing an Ethernet network must be approached with care. Many aspects of the network must be
considered before actual implementation can begin. A detailed plan must be laid out to ensure that all
the goals and obstacles are identified. In this chapter we will discuss many aspects of network design
by beginning with a single segment 10BASE5 Ethernet network and gradually building the single
segment into a large, multiple segment network. We will then discuss the design of a 10BASE2
network, a Fiber Optic network, and finally a 10BASE-T network.
10BASE5 Ethernet Network Design
NOTE
The characteristics and test requirements of Ethernet 10BASE5 cables
are presented in the Cabletron Systems Cabling Guide.
Single Segment 10BASE5 Ethernet Network
A single segment of thick Ethernet cable can be a maximum of 500 m in
length. The cable must be terminated at both ends with N-type connectors
and 50 ohm N-type terminators. The cable should be marked with annular
rings by the cable manufacturer every 2.5 m indicating potential
transceiver tap points (see Figure 7-1).
Annular Rings
Coaxial Cable
N-Type
Connector
2.5 m
(10BASE5)
1913-18
Figure 7-1. Coaxial Annular Rings
7-1
10BASE5 Ethernet Network Design
The coaxial cable can be run in one continuous length or in sections, joined
using N-type connectors and N-type barrel connectors. If the cable is
installed in segments and connected together, IEEE recommends that the
segments should be an odd multiple of 23.4 m in length. These special
lengths of cable are used to minimize signal reflections caused by the
insertion of connectors and barrel splices. The segment lengths are
optimum lengths and it is sometimes difficult to adhere to these
specifications in the real world. If it is not possible to adhere to the above
cable lengths, attempt to connect the cable segments at annular rings.
To further reduce signal reflections, the cable segments should come from
one manufacturer to minimize cable discontinuity problems such as
differences in propagation speed or impedance.
Transceiver Placement
The terminated coaxial cable is now ready for the placement of the
transceivers. Transceiver taps can be made into the coaxial backbone by
one of two ways:
1. Intrusive: By cutting the coaxial cable and installing connectors that
screw into the transceiver. This method of tapping is known as an
intrusive tap because the coaxial cable must be severed for the
transceiver to be connected.
2. Non-intrusive: By drilling a hole into the coaxial cable with a special
tool and clamping the transceiver assembly in place. This method is
known as a non-intrusive tap because the coring process, when done
properly, allows the tap to be made without interruption of network
traffic on the coaxial segment (see Figure 7-2).
Compression
Screw
Annular Ring
Coaxial Cable
Tap Pin
Transceiver
Contact
1913-19
Figure 7-2. A Non-intrusive Coaxial Cable Tap
7-2
Ethernet Network Design
10BASE5 Ethernet Network Design
A maximum length 10BASE5 coaxial cable has 199 annular rings marked
off at 2.5 m intervals. Because each transceiver tap introduces noise onto
the coaxial cable in the form of a small impedance discontinuity, and
contributes to the overall attenuation of the cable, IEEE has specified that
only 100 taps will be allowed on a 10BASE5 segment, with each tap
separated by a minimum of 2.5 m.
Once the coaxial taps are in place, they can be attached directly to the
transceivers (see Figure 7-3). An Attachment Unit Interface (AUI) cable is
used to connect devices such as workstations or file servers to the
transceivers. The AUI cable, which can be a maximum of 50 m in length, is
made up of four shielded twisted pairs that carry the transmit, receive and
collision signals between the transceiver and its connected equipment.
Coax Cable
Non-Intrusive Tap
Contact Pin
EIVER
SC
TRAN
AUI Connector
1913-20
Figure 7-3. Transceiver Attachment
Multi-port Transceivers
As you recall from the previous Chapter 6, Ethernet Network Design, a
multi-port transceiver is a device that is used to connect several network
devices at a single tap point. This is useful when it is necessary to
concentrate a number of devices at one physical location or to add more
than 100 devices on a single coaxial backbone. A typical multi-port
transceiver has eight AUI ports for connection to workstations and one
AUI port for connection to the backbone segment.
Ethernet Network Design
7-3
10BASE5 Ethernet Network Design
Coaxial Backbone
TRANSCEIVER
Multi-port Transceiver
AC
5
6
7
8
Network
22
22
22
4
22
3
22
2
22
22
22
22
1
To Network Devices
1913-21
Figure 7-4. Multi-port Transceiver
Multi-port Transceiver Rules
When multi-port transceivers are used, two rules must be observed:
1. Multi-port transceivers can be cascaded by connecting the male
connector of one multi-port transceiver to the female connector of a
second multi-port transceiver. However, cascading more than two
multi-port transceivers will result in unacceptable amounts of jitter,
causing alignment errors on the network to occur.
2. The maximum length of a standard AUI cable is 50 m. This maximum
length is reduced by 10 m for each multi-port transceiver that the
signal must pass through. In other words, if two multi-port
transceivers are cascaded, only 30 m total of AUI cable between them,
the device, and the coaxial tap, is allowed.
You can now see how it would be possible to construct a 10BASE5
Ethernet network that would support the IEEE maximum of 1,024 devices.
By using cascaded multi-port transceivers it is possible to have up to a
maximum of 64 user devices per coaxial tap point. See Figure 7-5.
7-4
Ethernet Network Design
10BASE5 Ethernet Network Design
Coaxial Backbone
TRANSCEIVER
Connect up to 8 MP
transceivers or devices
Accounts
for 10 m of
AUI cable
AC
2
3
4
5
6
7
22
22
1
8
Network
Connect up
to 8 devices
AC
1
2
3
4
5
6
7
2 2
Accounts
for 10 m of
AUI cable
8
Network
22
To Device
These cables add
up to a total of 30 meters.
1913-22
Figure 7-5. Cascaded Multi-port Transceivers
Grounding and Insulation
The only issue that remains before we have a functioning single segment
10BASE5 Ethernet network is grounding the coaxial cable and insulating
the connectors. The backbone cable must be connected to a reliable earth
ground at only one point. The actual connection to ground can be made at
any point on the cable, but is usually accomplished at an N-type
connector, which allows for convenient attachment to the coaxial shield
through the connector body.
All connections, other than the grounded connection, must be insulated
from any other metallic surface to avoid inadvertent grounding and
creation of ground loops (multiple paths to ground).
Once the connectors are insulated, the coaxial cable is grounded and the
workstations are connected to the transceivers, the single segment
network is ready for operation.
Ethernet Network Design
7-5
10BASE5 Ethernet Network Design
Multiple Segment 10BASE5 Ethernet Network
We have seen how we can build a single segment 10BASE5 Ethernet
network. This is adequate if we only want to span a distance of 500 m. If it
is necessary to cover a greater area or to add additional coaxial taps
beyond the 100 tap limit, more coaxial cable must be added. To connect the
new coaxial segment to the existing backbone, a repeater must be used.
As you recall from Chapter 6, Ethernet Devices, the repeater is a Physical
Layer device that has the capability to forward frames at up to full
Ethernet bandwidth. It regenerates the preamble and amplifies and
retimes the signal from one backbone cable to another. It is connected to
the coaxial cable through transceivers in the same fashion as any other
node on the network and requires external power to operate. The repeater
also extends fragmented frames and will auto-partition ports in the event
of excess collisions.
Repeater Use
When a local repeater is used to connect two 10BASE5 coaxial segments,
each segment may be a maximum of 500 m in length with up to 100 taps,
including taps made to connect the repeater. The maximum of 1,024
devices allowed and the AUI cable length limitation of 50 m apply to both
segments. It is not necessary that repeaters be connected at the ends of the
coaxial segment. Refer to Figure 7-6.
Network Segment 1
Network Segment 2
(Up to 1,024 devices)
(Up to 1,024 devices)
Repeater
1913-23
Figure 7-6. Repeater Use
7-6
Ethernet Network Design
10BASE5 Ethernet Network Design
Inter-Repeater Link (IRL)
Repeaters can be used to connect up to 3 coaxial segments before special
considerations need to be taken into account. If it is necessary to connect
more than 3 coaxial segments together, then an Inter-Repeater Link (IRL)
must be used (see Figure 7-7). An IRL is a segment that spans between two
repeaters with no other devices attached to it. It can be made up of thin
coaxial (185 m), thick coaxial (500 m), fiber optic cable (2/5 km) or twisted
pair cable (200 m) with the appropriate media limitations being observed.
IEEE states that if IRLs are used, the maximum network size can be up to 4
repeaters with 5 segments, giving a total linear distance of 2,500 m (if just
10BASE5 coaxial is used), not including AUI cables. IEEE has determined
that this is the maximum network size that will allow the round trip
propagation delay budget to be met. Propagation delay is discussed in
Chapter 8, Propagation Delay.
NETWORK 1
NETWORK 2
(Up to 1,024 devices)
(Up to 1,024 devices)
Inter Repeater Link
(length dependent
on media type)
Repeaters
1913-24
Figure 7-7. Using Inter-Repeater Links
NOTE
In multiple segment networks, each coaxial segment must be grounded at
only one point. All exposed metallic connectors and terminators must be
insulated from ground to prevent ground loops.
Ethernet Network Design
7-7
10BASE2 Ethernet Network Design
10BASE2 Ethernet Network Design
NOTE
The characteristics and test requirements of 10BASE2 cables are
presented in the Cabletron Cabling Guide.
Single Segment 10BASE2 Ethernet Network
10BASE2 is the IEEE specification for Ethernet running on RG58 A/U
coaxial cable. 10BASE2 coaxial cable is more flexible and less expensive
than 10BASE5 coaxial cable while still maintaining the required 50 ohm
nominal impedance. The maximum length of a 10BASE2 cable segment is
185 m. The reduced distance is due to the higher loss characteristics of the
RG58 A/U compared with 10BASE5 coaxial cable. Only 30 taps are
allowed on a 10BASE2 segment and the cable is connected using BNC
connectors and terminated using 50 ohm BNC terminators.
Tapping into a 10BASE2 cable is accomplished by cutting the cable and
attaching a BNC T-connector (see Figure 7-8). One end of the T-connector
is either attached to the network interface card in the workstation or to a
wall plate and the other two ends are attached to the coaxial cable
backbone. An AUI cable is used to go from the wall plate to the device or
workstation. The minimum spacing between T-connectors or splices is
0.5 m.
BNC T-Connector
1913-25
Figure 7-8. Typical BNC T-Connector
7-8
Ethernet Network Design
10BASE2 Ethernet Network Design
Workstation Connections
Workstations may be connected to the BNC T-connector in one of the
following two ways:
1. By connecting a transceiver with a BNC connection directly to the
T-connector then running up to 50 m of standard AUI cable to the
workstation.
2. By connecting the T-connector to the internal transceiver that is built
into most network interface cards on the device itself. This is the most
popular way to connect a device to the coaxial backbone in a 10BASE2
network.
NOTE
An AUI cable is used to go between the transceiver and the workstation,
not between the tap and the transceiver. If the internal transceiver on the
network interface card is used in a 10BASE2 network, the use of an AUI
cable is not allowed.
Grounding and Insulation
As with 10BASE5 networks, the 10BASE2 segment must be grounded at
only one point and all remaining connectors must be insulated from
contact with ground.
Multiple Segment 10BASE2 Ethernet Network
Since 10BASE2 has a physical limitation of only 30 devices and 185 meters
of coaxial cable, it is easy to see how a network might quickly grow
beyond specifications. As with 10BASE5 Ethernet, 10BASE2 can also be
expanded using repeaters. Although a standard 10BASE5 repeater and
BNC style transceivers may be used, the more common approach is to use
a multi-port repeater such as a Cabletron Systems MR9000C. The
MR9000C has eight BNC connections to allow the connection of up to 8
full specification 10BASE2 segments. A multi-port repeater counts as only
one repeater when calculating your maximum repeater path. Many also
have an AUI port to allow for connection to a standard transceiver on a
10BASE5 network.
Ethernet Network Design
7-9
10BASE2 Ethernet Network Design
By using repeaters, multi-port repeaters, or a combination of both, it is
possible to design a 10BASE2 network that has 4 multi-port repeaters and
five interconnected segments that span a distance of 5x185 m = 925 m,
with a maximum of 1,024 connected devices. As long as the longest
possible signal path does not pass through more than four repeaters and
five segments, with only three of the segments being populated with
devices, the network is within IEEE 802.3 specifications (see Figure 7-9).
ST-500 with Lanview
MP Repeater
weivnaL htiw 005-TS
MP Repeater
IRL
IRL
weivnaL htiw 005-TS
weivnaL htiw 005-TS
MP Repeater
MP Repeater
All Coaxial Segments are
185 m with 30 taps maximum
(except IRLs)
1913-26
Figure 7-9. Maximum Size 10BASE2 Ethernet Network
7-10
Ethernet Network Design
Fiber Optic Ethernet Network Design
Grounding and Insulation
When cascading multi-port repeaters be careful to avoid the creation of
ground loops (or multiple paths to ground). If two multi-port repeaters
that perform internal grounding are connected using the BNC ports, a
ground loop will result (see Figure 7-10). In other words, if one repeater is
connected using a BNC port, the other must be connected using the AUI
port.
ST-500 with Lanview
Improper Connection:
Causes Ground Loops
Proper Connection
1913-27
Figure 7-10. Multi-port Repeater Connection
Fiber Optic Ethernet Network Design
Communication over fiber optics is done with pulses of light transmitted
on glass instead of pulses of electricity transmitted on copper. Because of
the low loss and high noise immunity of fiber, it is the media of choice
when designing an extended distance LAN. A fiber optic segment can be
up to 5 km in length and is used as a point-to-point media where no taps
or branches are allowed.
Fiber optics is typically used in Ethernet networks to connect repeaters
between buildings and to traverse long distances as an IRL. It can also be
used with transceivers or fiber optic network interface cards to connect
network nodes.
A completely fiber optic network may be built as long as the 4 repeater, 5
segment and 1,024 devices rules are not broken. To connect multiple fiber
optic cable runs, a fiber optic multi-port repeater is used. Fiber is used as
an IRL to span between repeaters and multiple fiber drops to workstations
can be implemented (see Figure 7-11).
Ethernet Network Design
7-11
10BASE-T Twisted Pair Ethernet Network Design
Fiber Optic Inter-Repeater Link (FOIRL)
Fiber
concentrator
Fiber
concentrator
Fiber drop
to workstation
Fiber drop
to workstation
FOT
Fiber
transceiver
AUI
cable
1913-28
Figure 7-11. A Complete Fiber Optic Ethernet
10BASE-T Twisted Pair Ethernet Network Design
Unshielded Twisted Pair (UTP) wiring is found in most business
environments. For this reason, IEEE adopted a set of standards for
implementing this cost effective wiring into its own Ethernet category.
Designing a 10BASE-T Ethernet is much like designing a Fiber Optic
network, as both media are point-to-point media. A 10BASE-T network is
configured in such a way that the resulting topology is a star. A multi-port
repeater is used in a central location, such as a wiring closet, with twisted
pair segments going to the workstation locations. Each segment is a
maximum of 200 m in length, with no other taps or branches allowed. As
with all un-bridged Ethernet networks, the maximum signal path is 4
repeater hops and 5 segments. Refer to Figure 7-12.
7-12
Ethernet Network Design
10BASE-T Twisted Pair Ethernet Network Design
Twisted Pair
Repeater
Multiple Twisted
Pair Drops to
Workstations
1913-29
Figure 7-12. 10BASE-T Ethernet
Ethernet Network Design
7-13
10BASE-T Twisted Pair Ethernet Network Design
7-14
Ethernet Network Design
Chapter 8
Propagation Delay
From the preceding, Chapter 7, Ethernet Network Design, we have seen how it is possible to build a
maximum size network using each of the available media. When designing any Ethernet network, it is
wise to calculate the maximum round trip propagation delay for the proposed design. The maximum
allowable round-trip propagation delay of 51.2 µs, governed by the minimum frame size of 64 bytes, is
a very important consideration when it comes to accurate collision detection and data integrity.
Everything that lies in the signal path will contribute to the overall propagation delay. The items that add
delay are transceivers, repeaters, active hubs, passive hubs and cables. Bridges, which effectively
reset propagation delay, are not considered in the calculation of propagation delay. They will be
discussed in Chapter 9, Propagation Delay.
Calculating the Delay
As you design your network, you should sketch the overall topology to
determine rough equipment locations, cable segment loading and
maximum repeater hops. Calculating the overall round trip network
propagation delay is an easy task that should only take a few minutes. The
most difficult part, by far, is finding the proper delay times for individual
pieces of hardware. These numbers should be obtainable from the original
hardware manufacturer.
The total propagation delay is found by simply adding up the delays of
the individual components and cabling in the longest signal path and
multiplying the result by 2 to get the round trip delay time. The longest
signal path is found by identifying the longest path between any two
network devices, even though the two devices may not directly
communicate with each other. Examples of network devices that could be
considered an end point on the network would be PCs, workstations, file
servers, bridges, routers, gateways, and printers.
8-1
Propagation Delay Example
Propagation Delay Example
To clarify the methods used to calculate the propagation delay of a given
network, refer to Figure 8-1 and complete the following steps to calculate
the propagation delay of the network shown.
500 m 10BASE5
ST-500 with Lanview
ST-500 with Lanview
ST-500 with Lanview
5 m AUI
50 m
AUI
ST-500 with Lanview
5m
AUI
30 m AUI
Local
Repeater
185 m 10BASE2
Local
Repeater
5m
AUI
ST-500 with Lanview
C
45 m
AUI
50 m
AUI
Fiber
Transceiver
1000 m
Fiber Optic
File Server
Fiber Optic
Repeater
5m
AUI
500 m 10BASE5
500 m 10BASE5
ST-500 with Lanview
ST-500 with Lanview
ST-500 with Lanview
ST-500 with Lanview
ST-500 with Lanview
ST-500 with Lanview
5m
AUI
5m
AUI
50 m
AUI
Multi-port
Transceivers
10BASE-T
Concentrator
40 m
AUI
1m
AUI
B
A
100 m
Twisted
Pair
24 m
AUI
File Server
D
5m
AUI
10BASE-T
Transceiver
1913-30
Figure 8-1. Propagation Delay Example
1. Carefully examine Figure 8-1 and identify the longest signal path
between any two end devices. Remember, the longest path is not
necessarily the longest physical path, but the path with the longest
signal delay.
You should have determined that the longest path is the path between
Workstation B and Workstation D.
2. Using the equipment delay table, Table 8-1, make a list of the
equipment the signal must pass through in this path and the
appropriate delay time for each piece of equipment.
NOTE
8-2
The equipment delay times in Table 8-1 are Cabletron Systems specific
delay times. Other manufacturer’s equipment may have different delays.
Refer to the specific manufacturer’s delay times of the equipment in your
network design to calculate your propagation delay.
Propagation Delay
Propagation Delay Example
3. Now that you have the equipment delay times, you must calculate the
total delay for each media type. To do this, make a list of each media
type found in the signal path and add up the total length, in meters, of
each. Now, multiply this number by the appropriate media delay time
for each media type found in Table 8-1, to get the delay time for each
length of media found in the signal path.
NOTE
The values contained in Table 8-1 are maximum values. It is possible that
certain grades of cable will have a higher velocity of propagation which will
result in a smaller delay per meter.
4. Add up the equipment delay times, calculated in step 2, to find the
total equipment delay for this signal path.
5. Add up the media delay times, calculated in step 3, to find the total
media delay for this signal path.
6. Add the numbers found in step 4 and step 5. This is the total one-way
propagation delay time for this network. Multiply this number by two
to get the round trip propagation delay time.
By IEEE definition, the one way propagation delay time must be less than
or equal to 25.6 µs. Your result from this exercise should be 23.45 µs. Since
this value is less than the IEEE maximum, the network is within
specification. If you didn’t come up with 23.45 µs, refer to Table 8-2 and
Table 8-3 to see where you went wrong.
Propagation Delay
8-3
Propagation Delay Example
Table 8-1. Equipment and Cable Propagation Delay Times
8-4
Equipment
Type
Delay
Media Type
Delay per
Meter
Local Repeater
0.65 µs
10BASE5 coaxial
0.00433 µs/m
Fiber Optic
Repeater
1.55 µs
10BASE2 coaxial
0.00514 µs/m
Multi-port
Repeater
1.55 µs
Shielded Twisted
Pair (STP)
0.0057 µs/m
Multi-port
Transceiver
0.10 µs
Unshielded
Twisted Pair (UTP)
0.0057 µs/m
Standard
Transceiver
0.86 µs
Fiber Optic
0.005 µs/m
Fiber Optic
Transceiver
0.20 µs
AUI
Twisted Pair
Transceiver
0.27 µs
Concentrator
1.90 µs
0.00514 µs/m
Propagation Delay
Propagation Delay Example
Table 8-2. Equipment Propagation Delay Worksheet
Equipment Type
A
Delay
B
Quantity
(A*B)
Total
1
Local Repeater
0.65 µs
2
1.30 µs
2
Fiber Optic Repeater
1.55 µs
1
1.55 µs
3
Multi-port Repeater
1.55 µs
-
0
4
Multi-port Transceiver
0.10 µs
2
0.20 µs
5
Standard Transceiver
0.86 µs
6
5.16 µs
6
Fiber Optic
Transceiver
0.20 µs
1
0.20 µs
7
Twisted Pair
Transceiver
0.27 µs
1
0.27 µs
8
Concentrator
1.90 µs
1
1.90 µs
10.58 µs
Equipment Delay Total (Add lines 1–8)
Table 8-3. Cable Delay Worksheet
Cable Type
A
Delay/Meter
B
Quantity
(A*B)
Total
1
10BASE5
0.0043 µs/m
1500 m
6.5 µs
2
10BASE2
0.00514 µs/m
-
0
3
Shielded Twisted Pair
(STP)
0.0057 µs/m
-
0
4
Unshielded Twisted
Pair (UTP)
0.0057 µs/m
100 m
0.57 µs
5
Fiber Optic
0.005 µs/m
1000 m
5.00 µs
6
AUI
0.00514 µs/m
155 m
0.80 µs
Cable Delay Total (Add lines 1–6)
Propagation Delay
12.87 µs
8-5
Propagation Delay Example
To calculate the One Way propagation delay time, add the values in the
shaded bottom right hand corner of Table 8-2 and Table 8-3 together.
Equipment Delay:
Cable Delay
10.58 µs
12.87 µs
Total One Way Delay
23.45 µs
As with designing anything, it is not advisable to design to the limits of
the specification. As an electronic device ages, its internal propagation
delay may change. Also, as devices such as repeaters and transceivers are
changed due to upgrades or changing from one vendor to another, the
propagation delay times may change as well. By using the worst case
values for the delay times you should have a comfortable amount of
design buffer built into your network.
8-6
Propagation Delay
Chapter 9
Ethernet Bridge Operation
Bridges are devices that are added to a network to allow expansion beyond the limits of the IEEE
802.3 specification. They are also used to connect two similar networks together, allowing
communication between them. Bridges accomplish this by reading in frames and deciding to either
filter or forward the frame based on the destination address of the frame. The following sections detail
the operation of bridges and their functions.
Filtering and Forwarding
Bridges decide whether to forward or filter a frame based on the location
of the destination with respect to the source. They dynamically learn the
location of devices by logging the source address of each frame and the
port it was received on in their Source Address Table (SAT). Refer to
Figure 9-1, which shows two Ethernet LANs connected by a bridge, and
the following explanation on the filtering and forwarding process.
Network 2
Network 1
ST-500 with Lanview
ST-500 with Lanview
ST-500 with Lanview
weivnaL htiw 005-TS
BRIDGE
A
Station A
B
Station B
C
Station C
D
Station D
1913-31
Figure 9-1. Use of Ethernet Bridges
9-1
Filtering and Forwarding
When the bridge is first powered up, its SAT is empty:
Network 1
Network 2
Assume Station A wants to transmit a frame to Station B. The bridge
receives the frame and checks the CRC (Cyclic Redundancy Check) of the
frame. The bridge then looks at the source address of the frame and puts
that address in the source address table of Network 1 as shown below:
Network 1
Network 2
A
The bridge then checks the destination of the frame to see if it is located on
Network 1. Since Station B has not transmitted anything yet, the bridge
has no idea where it is located so the bridge forwards the frame to
Network 2.
Station B now sends its response to Station A. The bridge receives the
frame and checks the CRC. If the CRC is good, the bridge looks at the
source address and sees that the address is B. The Station B address is
placed in the SAT as shown below:
Network 1
Network 2
A
B
Next, the bridge checks to see if the destination address (Station A) is
listed in the SAT. In this case, Station A is listed as being on Network 1 so
the bridge blocks the frame from Network 2.
9-2
Ethernet Bridge Operation
Spanning Tree Algorithm
Assume that station A wants to transmit to Station D. Station A sends the
frame to Station D. The bridge reads in the frame and checks the CRC. The
bridge reads the source address of the frame (Station A) and makes sure
Station A is still in the SAT. The bridge checks the SAT for the destination
address, Station D. It is not found to reside on the Network 1 side of the
bridge so the frame is forwarded to Network 2.
Next, Station D sends its response to Station A. The bridge reads in the
frame, and after checking the CRC, updates the SAT with the Station D
address as being located on Network 2. The SAT now shows the
following:
Network 1
Network 2
A
D
B
The bridge then inspects the SAT looking for Station A on the Network 2
side of the bridge. It was not found on the Network 2 side so the frame is
forwarded to Network 1.
NOTE
It is important to realize that the bridge did not make the forwarding
decision because Station A was on the Network 1 side, but because
Station A was not on the Network 2 side.
This process continues until all stations are logged on the SAT. The bridge
will now isolate different network traffic, as well as extend the maximum
size of each individual network.
Spanning Tree Algorithm
Considering the important role bridges play in the transfer of data from
one network to another, it is a good idea to set up a redundant bridge that
commences operation if the primary bridge should fail. Therefore, IEEE
chose to build some fault tolerance into the bridge specification. The
802.1d specification defines bridge operation, redundancy and a process
called the Spanning Tree Algorithm (STA) which allows bridges to be
connected in such a way as to create standby redundant paths without
creating data loops. This same algorithm activates a redundant path in
case of a failure in the active path.
Ethernet Bridge Operation
9-3
Spanning Tree Algorithm
Configuration BPDU
When a bridge is powered up, it goes through a series of self tests to check
its internal operation. During this time the bridge is in a standby condition
and does not forward traffic. Also during this standby period, the bridge
sends out special bridge management frames called Configuration Bridge
Protocol Data Units (BPDU). Bridges use the BPDU frames as a way of
communicating with each other. The configuration BPDU is 64 bytes long
and contains the following fields:
9-4
•
Destination Address: A specific 6-byte Ethernet multicast address that
denotes the bridge group.
•
Source Address: The standard 6-byte Ethernet hardware address of the
bridge transmitting the BPDU.
•
Length Field: A standard, 2-byte, IEEE 802.3 data field length.
•
Data Field: Contains 35 bytes for BPDU data and 13 bytes of pad to
equal the minimum data field size of 48 bytes. The BPDU data field
contains the following information:
-
Protocol Identifier: A reserved, 2-byte, protocol identifier defined
by IEEE.
-
Protocol Version Identifier: Identifies the version of the bridge
protocol being used and is 1 byte in length.
-
BPDU Type: 1 byte to identify the BPDU type as either a
configuration or topology change BPDU.
-
Flags: A 1-byte field that contains topology change and topology
change acknowledgment flags.
-
Root Identifier: An 8-byte identification number derived from the
Ethernet address of a bridge and its unique port addresses. This
field contains the ID for the perceived root bridge.
-
Root Path Cost: A 4-byte field that contains a “cost” value
composed of individual bridge port costs along a data path. Bridges
use this information to determine the optimum frame transmission
path.
-
Bridge Identifier: An 8-byte identification number that is derived
from the Ethernet address of a bridge and its unique port addresses.
-
Port Identifier: A 2-byte field that contains port priority information
based on unique bridge port addresses. The port with the lowest
address has the highest priority.
Ethernet Bridge Operation
Spanning Tree Algorithm
-
Message Age: A 2-byte field that contains the age of the
configuration BPDU. This parameter allows a bridge to determine
if the BPDU is too old and needs to be discarded.
-
Max Age: A 2-byte field that contains a time-out value initially set
by the root bridge. This value is compared to the Message Age to
determine the validity of the BPDU.
-
Hello Time: A 2-byte field that contains the value for the time
interval used to generate configuration BPDU’s by the root bridge.
If a bridge does not hear from the root bridge within the time
defined by the Hello Time, the bridge will initiate a topology
change and attempt to become the new root.
-
Forward Delay: A 2-byte field that contains a value used by all
bridges as a delay value when a port changes state to the
forwarding condition. This delay is necessary to prevent data loops
and duplicate data that could be caused by an instantaneous
change of port state. This delay value is also used as the age time for
the source address table whenever a topology change has been
detected by a bridge. This value temporarily replaces the default
age time in order to quickly flush the source address table so
network addresses will be properly relearned after the topology
change. The Forward Delay time value is established by the root
bridge.
The purpose of the configuration BPDU is to notify other bridges on all of
the connected networks of the current topology. Based on the bridge
priority and address, the other bridges automatically detect loops and
negotiate a single path. The bridge or bridges involved in this primary
data path then come on-line and the bridges with lower priority involved
in the backup path(s) remain in standby.
Topology Change BPDU
A topology change BPDU is made up of only 4 data bytes (plus pad) that
contain the Protocol Identifier, Protocol Version Identifier and BPDU Type.
The purpose of the topology change BPDU is to notify other bridges that a
change has taken place. The other bridges then re-span to form a legal
topology.
Ethernet Bridge Operation
9-5
Spanning Tree Operation
Spanning Tree Operation
In the following explanation we take Spanning Tree through the network
shown in Figure 9-2, which consists of two-port bridges. You must
understand that the Spanning Tree process is a single operation,
combining both root bridge determination and data loop detection and
resolution. For the purpose of explanation, we split the process into two
individual discussions:
•
Root Bridge determination
•
Data Loop detection and resolution
Also, to clarify our referencing, the bridges are named as shown:
LAN A
ANN
JANET
LAN B
SAM
EVENIN
LAN C
1913-32
Figure 9-2. Network Before Spanning Tree
The Bridge Parameters for the network above are:
9-6
Bridge
Bridge ID
Path Cost
SAM
80-00-00-00-1d-23-56-a2
100
ANN
80-00-00-00-1d-56-d4-f4
100
JANET
80-00-00-00-1d-f4-67-2a
100
EVENIN
80-00-00-00-1d-4f-94-a1
100
Ethernet Bridge Operation
Spanning Tree Operation
The primary function of the Spanning Tree Algorithm is to ensure that
there is only one data path between any two end stations within the
bridged Local Area Network. All computations are geared towards the
fact that a bridge wants to be considered as the Designated Bridge for any
LAN that it is connected to. Upon power up, BPDUs are sent out as
multicast frames. A bridge directly connected to the same LAN that the
BPDUs are sent out on will accept these BPDUs and make decisions on
their contents.
For ease of explanation we will represent the individual bridge’s BPDUs
in a shortened version consisting of four parameters as shown below:
BPDU
Root Bridge ID
Path Cost
Bridge ID
Port ID
•
Bridge ID: A 64-bit value that is comprised of two individual
components:
-
Bridge Priority: A 16-bit configured value
-
MAC Address: A 48-bit value which is the hardware Ethernet
address of the bridge, e.g., 00-00-1d-e2-45-a2
•
Root Bridge ID: Same as the Bridge ID but is the ID of the designated
root bridge.
•
Path Cost: A value representing the contributing cost of passing
through this bridge. The formula used to determine the default is
1000/network Mbps per sec. The default for Ethernet is 1000/10=100.
•
Port ID: A 16-bit value made up of two components:
-
Port Priority: 8 bits in length and is the most significant byte of the
Port ID.
-
Port ID: 8 bits in length numbered sequentially on a bridge from 1
to infinity (in theory).
The default Port ID for port 1 resembles the following: 8000-0001.
Ethernet Bridge Operation
9-7
Spanning Tree Operation
With the understanding of our shortened BPDU, we now begin our
explanation of the Spanning Tree Operation. In Figure 9-3, SAM is sending
out BPDUs across all LANs to which it is connected. SAM sends BPDUs
out through its port 2 onto LAN C and through port 1 onto LAN A.
LAN A
ANN
JANET
Port 1
LAN B
SAM
Port 2
EVENIN
LAN C
SAM's BPDUs
1913-33
Figure 9-3. SAM’s Initial BPDUs
SAM’s BPDUs are represented by the following examples:
BPDU (Port 1)
BPDU (Port 2)
80-00-00-00-1d-23-56-a2
80-00-00-00-1d-23-56-a2
0
0
Bridge
ID
80-00-00-00-1d-23-56-a2
80-00-00-00-1d-23-56-a2
Port ID
8001
8002
Root ID
Path Cost
Each of these BPDUs represent SAM as their root bridge. This is shown by
the Path Cost to the root from LANs A & C = 0 (if SAM was root, then
there would be no bridge hops going from LAN A to the root, hence a cost
of 0). They also indicate that SAM was the transmitting bridge (Bridge ID).
The only difference between the BPDUs is the Port Identifier; 8001 for port
1 and 8002 for port 2.
9-8
Ethernet Bridge Operation
Spanning Tree Operation
Now let’s look at one of the other bridges, in this case ANN. Figure 9-4
shows ANN generating BPDUs from all of its ports, considering itself as
root until it finds out differently.
LAN A
Port 1
ANN
JANET
Port 2
LAN B
SAM
EVENIN
LAN C
ANN's BPDUs
1913-34
Figure 9-4. ANN’s Initial BPDUs
The BPDUs generated by ANN are shown below:
BPDU (Port 1)
BPDU (Port 2)
80-00-00-00-1d-56-d4-f4
80-00-00-00-1d-56-d4-f4
Path Cost
0
0
Bridge ID
80-00-00-00-1d-56-d4-f4
80-00-00-00-1d-56-d4-f4
8001
8002
Root ID
Port ID
Here again, this bridge thinks that it is the root, indicated by the root ID
value within the BPDU, and the cost to the root of 0. It also indicates by the
Bridge ID that ANN is the bridge that has transmitted this BPDU. Again,
the Port Identifier reflects the transmitting port of the BPDU on ANN. We
now assume that both SAM and ANN receive each other’s BPDUs from
LAN A.
Ethernet Bridge Operation
9-9
Spanning Tree Operation
SAM checks the incoming BPDU from ANN and it reflects a different root
ID. In the process of checking the incoming data against that which is
current at that port, SAM realizes that it has the higher priority root ID
(lower number) and does not forward ANN’s BPDU through port 2; it
continues to send its own.
ANN checks the incoming BPDU from SAM and senses that the BPDU
carries a higher priority BPDU (lower number) than its own. ANN now
stops transmitting its own BPDUs and begins to modify and retransmit
those received from SAM. These retransmissions by ANN are transmitted
out of port 2, but not port 1. This procedure continues until the only
BPDUs being generated are originating at the root bridge.
Before we jump to the root bridge, however, let’s continue the BPDU trail
to the finish. ANN relinquished its bid to become root and agreed that
SAM is better qualified. ANN propagates SAM’s BPDUs to any other LAN
that it is directly connected to. Here are ANN’s new BPDUs:
BPDU IN (Port 1)
BPDU OUT (Port 2)
80-00-00-00-1d-23-56-a2
80-00-00-00-1d-23-56-a2
Path Cost
0
100
Bridge ID
80-00-00-00-1d-23-56-a2
80-00-00-00-1d-56-d4-f4
8001
8002
Root ID
Port ID
You should recognize the BPDU coming in on port 1 (SAM’s BPDU). Now
look at the BPDU coming out of port 2. It indicates that SAM is the root
bridge (root ID) and the Bridge Identifier indicates that this BPDU was
transmitted from ANN. Notice that the root Path Cost is updated to
100–meaning it costs 100 to get to the root through ANN.
9-10
Ethernet Bridge Operation
Spanning Tree Operation
A similar case can be made for EVENIN, the bridge that spans LAN B and
LAN C shown in Figure 9-5.
LAN A
ANN
JANET
LAN B
SAM
BPDU in
from ANN
Port 1
EVENIN
BPDU out
Port 2
LAN C
1913-35
Figure 9-5. BPDUs in from ANN to EVENIN
EVENIN would have also thought that it was root until it found out
otherwise. Let’s look at the progression of BPDUs from ANN. Here are
EVENIN’s BPDUs:
BPDU IN (Port 1)
BPDU OUT (Port 2)
80-00-00-00-1d-23-56-a2
80-00-00-00-1d-23-56-a2
Path Cost
100
200
Bridge ID
80-00-00-00-1d-56-d4-f4
80-00-00-00-1d-4f-94-a1
8002
8002
Root ID
Port ID
The BPDU coming in to port 1 on EVENIN indicates that SAM has a
higher priority than it does, so it also stops generating its own BPDUs and
forwards those of the higher priority. The BPDU coming out of port 2 of
EVENIN reflects SAM as the root (root ID). The incoming BPDU reflects
that ANN can get to the root from LAN B at a cost of 100. The outgoing
BPDU onto LAN C represents the fact to any Bridge on LAN C that
EVENIN can get to the root for a cost of 200 (root path cost).
Ethernet Bridge Operation
9-11
Spanning Tree Operation
Data Loop Resolution
The center of our discussion here will be within EVENIN. It received
BPDUs from both of its ports. A BPDU from SAM came in on port 2, and a
BPDU from ANN came in on port 1. With the reception of these two
frames from different ports, each identifying SAM as the root, EVENIN
realizes that there is a data loop present.
Looking at this objectively, we can see that the root bridge is directly
connected to both LANs A and C so the root bridge is the designated
bridge for both of these LANs. However, the designated bridge for LAN B
is still undecided.
The BPDU transmitted from EVENIN onto LAN B and the BPDU
transmitted from ANN onto LAN B are shown next:
EVENINs BPDU OUT
(port 1) to ANN
ANNs BPDU OUT
(port 2) to EVENIN
80-00-00-1d-4f-94-a1
80-00-00-00-1d-23-56-a2
100
100
Bridge
ID
80-00-00-00-1d-4f-94-a1
80-00-00-00-1d-56-d4-f4
port ID
8001
8001
Root ID
Path
Cost
At this point we see that SAM is considered to be the root bridge by all
other bridges (root ID). There is no conflict concerning which bridge is the
designated bridge for LANs A and C, as the root is directly connected to
these two LANs. However we have two bridges trying to service LAN B:
ANN and EVENIN. This problem is resolved by the Bridge Entities
residing in both of these bridges.
Each bridge entity maintains a list of current port parameters for each
bridge port. The bridge entity receives BPDUs coming in from a port and
compare the BPDU information against the current port parameters. In the
case of a data loop, the bridge entity makes a decision as to whether it
should put the receiving port in the BLOCKING state or FORWARDING
state. It is this periodic dialogue that eventually settles our Bridged Local
Area Network in a single Spanning Tree.
In this scenario ANN retransmits SAM’s BPDUs onto LAN B and EVENIN
retransmits SAM’s BPDUs onto LAN B.
9-12
Ethernet Bridge Operation
Spanning Tree Operation
The information that both the bridges have in relation to LAN B is as
follows:
Port Parameters
ANNs
Port 2
EVENINs
Port 1
Port ID
8002
8001
Designated Root
SAM
SAM
Designated Bridge
ANN
EVENIN
Designated Port
8002
8001
Root Path Cost
100
100
The algorithm to determine who breaks the identified data loop is in the
following order:
a. lowest root path cost
b. highest priority designated bridge ID
c. highest priority designated port ID
d. highest priority port ID
Each bridge compares the incoming BPDU data up against its current port
parameters, and determines the active topology based upon the algorithm
above.
1. The root path cost is the first parameter checked. Both ANN and
EVENIN offer a cost of 100 from LAN B to the root so this is a
deadlock.
2. The designated bridge parameter is the next parameter checked:
-
ANN’s designated bridge is itself: 80-00-00-00-1d-56-d4-f4
-
EVENIN’s designated bridge is itself: 80-00-00-00-1d-4f-94-a1
Here, it appears that we have broken the tie. The bridge that holds the
highest priority designated bridge ID will win out. EVENIN sees an
inferior bridge ID coming in compared to its current parameters, so it will
go to the FORWARDING state. ANN sees a superior bridge ID coming in
so it will go to the BLOCKING State. EVENIN becomes the designated
bridge to LAN B.
Ethernet Bridge Operation
9-13
Spanning Tree Operation
There is one final data loop present. JANET is bridging from LAN A to
LAN B. The bridge entities recognize the data loop condition by
monitoring the incoming BPDUs. Upon seeing BPDUs coming in through
both of its ports, all originating from the root, it realizes that there is more
than one path to the root bridge. By using the order of comparisons shown
above, the bridge entity can make an intelligent decision as to whether the
port in question should be made part of the active topology or should be
sent to the BLOCKING state.
Figure 9-6 shows the resulting topology. There is only one data path
between any two end stations within this bridged LAN. All data loops
have been identified and resolved. We now have built-in redundancy that
can be used if one or more of the active bridge components fail.
LAN A
Port 1
Port 1
ANN
Port 1
JANET
Port 2 Blocking
Port 2 Blocking
LAN B
SAM
Port 2
Port 1
EVENIN
Port 2
LAN C
1913-36
Figure 9-6. Resulting Topology after Spanning Tree
9-14
Ethernet Bridge Operation
Index
Numerics
D
10BASE2
network design 7-8
10BASE5
network design 7-1
10BASE-T
network design 7-12
Data Field 4-4
Data Frame Type 4-5
802.2 4-7
Ethernet II 4-6
Raw 4-6
SNAP 4-9
Destination Address 4-3
A
Addressing
broadcast 4-11
multicast 4-11
specific 4-10
Annular Rings 7-1
Attachment Unit Interface 3-5
Auto Partition 6-4
E
Ethernet
data frames 4-3
devices 6-1
features 2-2
history 2-1
standards 3-1
B
F
Backoff 2-2
Bandwidth 2-2
BPDU 9-4
BPDU Type field 9-4
Bridge Identifier 9-4
Bridge Operation 9-1
Bridges 6-5
Fiber Optic Cable
network design 7-11
Filtering and Forwarding 9-1
Flags 9-4
Forward Delay 9-5
Frame Transmission 2-4
C
Calculating Propagation Delay 8-1
Cascade 7-5
Clean Packet Transmission 5-1
Collision 5-2
Collision Detection
point-to-point 5-3
Configuration BPDU 9-4
Contention Star Topology 2-7
Cyclic Redundancy Check 4-4
G
Ground Loops 7-11
Grounding and Insulation 7-5, 7-9
H
Hello Time 9-5
I
Inter-Repeater Link 6-5, 7-7
Intrusive Tap 7-2
Index-1
Index
L
R
Length Field 4-3
Logical Link Control 3-4
Repeater 6-3
Root Identifier 9-4
Root Path Cost 9-4
Routers 6-6
M
Manchester Encoding 4-1
Max Age 9-5
Media Access Control 3-4
Media Access Method 2-2, 5-1
Medium Dependant Interface 3-5
Message Age 9-5
Multi-point 2-5
Multi-port Repeaters 6-5
Multi-port Transceivers 6-3
Signal Quality Error 5-2
Source Address 4-3
Source Address Table 9-1
Spanning Tree Algorithm 9-3
Spanning Tree Operation 9-6
Start Frame Delimiter 4-3
Stations 6-1
N
T
Network Design 7-1
Non-intrusive Tap 7-2
T-connector 7-8
Topologies 2-4
bus topology 2-5
hybrid topology 2-7
ring topology 2-6
star topology 2-6
Topology Change BPDU 9-5
Transceiver 6-2
Transceiver Rules 7-4
O
OSI Model 3-1
application layer 3-2
data link layer 3-4
network layer 3-3
physical layer 3-5
presentation layer 3-3
session layer 3-3
transport layer 3-3
Out of Window Collision 5-3
P
Physical Layer Signaling 3-5
Physical Medium Attachment 3-5
Point-to-Point 2-5
Polling Star Topology 2-6
Port Identifier 9-4
Preamble 4-3
Propagation Delay 8-1
Protocol Identifier 9-4
Index-2
S