Operating System Security Rules

Operating System Security Rules
SECTION
9
Operating System Security Rules
This section addresses best practices for setting up security within
operating systems. Authentication, file protection, virus checking, file
sharing, network software, and security logging are discussed.
9.1 TRUSTED OPERATING SYSTEMS
Trusted operating systems have security features built into the operating
system. The National Computer Security Center’s Rainbow series
Orange Book, Trusted Computer Standards Evaluation Criteria describes several levels of trust including C1, C2, B1, B2, and B3.
Currently, there are no commercial operating systems that have been
certified beyond B1. These B1 operating systems are used primarily by
the government. To secure systems with any level of confidence, operating systems that are capable of C2 or B1 security should be used in your
network environment.
9.1.1 B1 Trusted Operating Systems
B1 level trusted operating systems are considerably more expensive
than C2 level operating systems because they are used in very selective
markets. As the commercial segment begins to deploy these operating
67
Information Security Best Practices – 205 Basic Rules
systems in greater numbers, the price should come down. Currently, B1
operating systems should be used for government and military applications that need to transmit sensitive, but not secret, data and for
commercial banking, financial services, and other commercial applications where security and confidentiality are very important.
9.2 AUTHENTICATION
✔
INFOSEC Best Practice #68
All server operating systems must have a minimum of a C2
level of trust.
C2 is a government designation for computer security that requires
computers to have the ability to control access to the computer via
usernames and passwords, and protect files by assigning ownership and
access rights. Desktop operating systems are recommended to have at
least C2 compliance to protect individual systems from unauthorized
access. Without these features the data on a computer system is vulnerable to anyone having access to the physical system. Some versions of
operating systems have most of the C2 features yet are not C2 compliant. This may be acceptable upon review of the missing features. Most
systems such as UNIX and NT can be configured to meet C2 compliance. C2, however, is somewhat dated and has many limitations.
Securing most server based operating systems will require going beyond C2 compliance with configuration and protection as specified in
the Best Practices outlined in this manual. The government also has
68
Operating System Security Rules
more stringent security designations such as B1, B2, and B3. Missioncritical server operating systems may need to have B2 compliance.
These security levels address the carrying of classification tags on all
data. Data objects are tagged with a classification level such as secret,
classified, confidential and unclassified. Access to these data objects is
available only to other objects that carry the correct classification level.
This type of classification scheme may also be applied to corporate data
with differing levels of sensitivity.
✔
INFOSEC Best Practice #69
Username and password authentication must be used to
access desktop computers.
Username and password authentication must be used to logon to
server and desktop computer systems (Reference Green Book “Department of Defense Password Management Guideline” CSC-STD-002-85).
Installing an operating system with this access control capability will
allow the user to secure the computer when not in use and prevent
illegal entry onto the system.
✔
INFOSEC Best Practice #70
Use one-time password generation hardware or software
to authenticate users for access to systems with missioncritical data.
69
Information Security Best Practices – 205 Basic Rules
Systems that have sensitive corporate data must authenticate users
with one-time passwords generated by handheld devices or software
such as S/Key. S/Key requires that the remote host know a password
that will not be transmitted over an insecure channel. When you connect
to the host, you get a challenge response. The challenge information and
password that you know plug into an algorithm that generates the appropriate response to the challenge from the remote host. The
algorithm-generated response should match the challenge response if
the password is the same on both sides. Thus, the password is never sent
over the network, and the challenge is never used twice.
Hardware authenticators should be PIN/Synchronous and require
keying in a PIN by the user to the authenticator device that then produces a password as a function of its internal clock. The user then enters
the generated password into the computer system. This type of system
authenticates the user in possession of the handheld authenticator and
limits eavesdropping of passwords to non-network means. It increases
security by requiring the user to know a password and additionally to
carry an authenticator that usually has a limited distribution. Handheld
authenticators should be kept in a secure area and issued to people only
when they need to use the system and then returned after use. As an
alternative to the handheld authenticator, a smart card can be used
which requires entering a PIN into the smart card followed by generation of a one-time password. Interaction with the computer takes place
via a smart card reader that doesn’t require the user to enter the generated password. This may be done at initial login only, if multiple
personnel use the machine.
70
Operating System Security Rules
✔
INFOSEC Best Practice #71
Use biometric authentication for systems that hold highly
sensitive information.
Some systems with very sensitive or government classified data may
require the use of a biometric device as the first step in a multilevel authentication process. A biometric device requires authentication of the
user by scanning a physical attribute as proof of identity. Examples of
biometric devices are fingerprint, retina scanners, handprint, and facial
feature devices. These devices are all coming down in price, which make
them more attractive for a variety of applications. This type of authentication requires the individual’s presence and cannot be circumvented as
easily as stolen smart cards or passwords.
As an example, a smart card can hold information about a person
such as fingerprint, photo, and other identifying information. When a
user desires access to a closed area or to a computer, a smart card is
scanned, the fingerprint is scanned, and compared to what is on the
card. If there is a match, then access is granted. Photos and other descriptive information are useful when there is a human guard that
provides another level of check. This is useful for very secure environments such as airports.
✔
INFOSEC Best Practice #72
Expire passwords after a designated period of time.
71
Information Security Best Practices – 205 Basic Rules
Your corporate security policy should specify a schedule indicating
how often each user’s password must be changed. This schedule must be
incorporated into the security policy for user accounts so that all user
account passwords expire after a designated period. It is very important
that user accounts not have a “never expire” flag set. Users do not want to
change passwords often and system administrators do not want to be
bothered about password expiration problems. Therefore, setting a sensible expiration password expiration policy increases the security of your
system. Examples of periods often used for password expiration are 60
days and 90 days. The policy for expiration must be linked to the type of
environment within an organization. For example, if there is a large
turnover of personnel or temporary accounts that are granted on the
system, then the expiration of passwords should be shorter as compared to
systems with relatively few accounts and no organizational turnover.
✔
INFOSEC Best Practice #73
Use “pass phrases” or a mix of letters and numbers for
passwords.
Passwords must NOT include the following:
a.
a portion or variation of an account username,
b. a portion or variation of your real name, address or phone
number,
c.
words or variations of words in a dictionary,
d. words or variations of foreign words,
72
Operating System Security Rules
e.
words spelled backward,
f.
repeated character strings,
g. only numeric digits,
h. only alphabetic characters,
i.
passwords less than six alphanumeric characters.
The weakest point in security is usually the password. People choose
short, easy-to-remember passwords that they do not change often.
Passwords must ideally consist of a mix of characters and numbers
interspersed in a long cryptic sequence (i.e., k3k5k*&eklx!!!i0). This is
too difficult to remember, therefore using a long phrase that is meaningful to you personally is a good alternative. This may be a code which
has meaning for a user, such as “My real age is 43” or “I have 6 pointyhaired bosses.”
✔
INFOSEC Best Practice #74
Check passwords against a password history file.
Sophisticated operating systems such as UNIX and Microsoft NT
contain a user account option for maintaining a history of used passwords. The program will not let a user reuse an old password for a
designated period of time. Therefore, the user must use new passwords
that nobody else will know or remember.
73
Information Security Best Practices – 205 Basic Rules
✔
INFOSEC Best Practice #75
Enforce a minimum password length of at least six
alphanumeric characters.
Most user account management programs allow the administrator to
enforce a minimum length for a password. The longer the password, the
less chance that a password cracking program will find your password.
Most password cracking programs generate combinations of letters and
numbers to create a password that is used to try to break into an account. The longer the password and the more characters that are used,
the more difficult it is to crack it. A seven-letter word has 247 combinations, if letters are used, and 347 combinations if both letters and
numbers are used. Each of those passwords has to be tried against an
account on the system. However, it is possible that the password program will generate your password early in the password generation
process. Using a phrase for your password makes it difficult to crack
your password.
✔
INFOSEC Best Practice #76
Automatically disable the user account if there are more
than six bad login attempts.
An attack on a user account will use a program that generates passwords one at a time against a user account until it successfully logs into
the system. This usually requires many passwords to be tried, unless a
common password is used. Most operating systems have an account
74
Operating System Security Rules
security feature that can be set by the system administrator to automatically disable an account after a designated number of login attempts.
This flag is often set to disable the account after about six unsuccessful
login attempts. It is highly unlikely that the user will enter their password incorrectly that many times. When the account is disabled only the
administrator will be able to enable the account again. If an attack is
actually occurring, the administrator must change the user name of the
account and make sure the user selects a satisfactory password. A message must be sent to all users on the system warning users that attacks
are occurring on the machine and that their passwords may have to be
changed to comply with security guidelines.
✔
INFOSEC Best Practice #77
Use Kerberos authentication in highly vulnerable
environments.
Kerberos is a network security protocol that uses strong encryption
and a complex ticket-granting algorithm to authenticate users on a network and to allow encrypted data streams over an IP network. Some TCP/
IP communication protocols such as FTP and TELNET transmit user IDs
and passwords in clear text. Kerberos is a good solution in environments
such as universities where other security mechanisms are difficult to
implement and enforce and where the environment needs to be open. It is
difficult to set up, but is becoming more ubiquitous and will be distributed
in conventional operating systems.
75
Information Security Best Practices – 205 Basic Rules
9.3 ACCOUNT SECURITY
✔
INFOSEC Best Practice #78
Do not disclose a computer’s identity until login is
completed successfully.
Set up the operating system so that the system login screen does not
identify the computer system by name or function until after login is
complete. Unauthorized personnel do not need to know the identity of
machines unless they need to use them. Hackers find this type of information valuable since it may identify valuable targets to break into.
✔
INFOSEC Best Practice #79
Use an automatic password generator to help the user
with password creation.
The more constraints your security policy puts on creating an adequate password, the more trouble users will have in creating
passwords. This may lead to frustration and complaints. To help the user
in choosing a password, some account management programs have an
automatic password generator that will produce a password according to
your security policy criteria. The drawback to using a password generator is that it can create cryptic passwords that may be difficult to
remember. Enabling this option, however, will at least provide some
help to frustrated users. Emphasis should be put on choosing your own
creative passwords that you can more readily remember.
76
Operating System Security Rules
✔
INFOSEC Best Practice #80
The password file must be encrypted by the operating
system.
Passwords are encrypted by most operating systems. If you are using
an old version of an operating system that does not encrypt passwords,
upgrade to a newer version of the operating system or modify the login
procedure to run your own program which reads a file encrypted by
your own method. Storing passwords in a file on your system that is
readable by everyone gives you no security and WILL lead to unauthorized access to your system(s). The password file should be modified
only by the system administrator.
✔
INFOSEC Best Practice #81
Restrict access by time of day for high-risk users.
Some users such as part-time employees, temps and consultants may
need to be given restricted access to a system only during business
hours because they may be considered to be more risky employees.
Most operating systems allow the system administrator to set time
restrictions on an account based on time of day.
✔
INFOSEC Best Practice #82
Audit all login attempts and set alarms for repeated
incorrect logins.
77
Information Security Best Practices – 205 Basic Rules
Keep security logging turned on and check the security log daily.
The log will be able to tell the system administrator what is happening
on the system and network. Suspicious activity may require more detailed monitoring with a sniffer, increased security protections, and
notification of users. This may prevent or diminish the extent of damage
if there is an attack on your system.
✔
INFOSEC Best Practice #83
Encrypt passwords that are transmitted over a network.
Authentication of users over a network typically involves sending a
password to a file server in clear-text. Therefore, the password is detectable by any sniffer monitoring the network. When choosing an
operating system be aware of this problem. Depending on your security
policy, you may have to provide third-party mechanisms that encrypt
passwords during authentication (e.g., Kerberos authentication, RSA
MD4).
✔
INFOSEC Best Practice #84
Create an alternate administrator or system account; do
not use the default.
The system account is one of the accounts that is typically attacked
by a hacker. Disabling interactive access to it and using another account
78
Operating System Security Rules
as the administration account will force the hacker to find another
legitimate username. Do not delete the system account since it may be
needed by various system functions. Finding a legitimate username may
not be difficult because most companies use very similar username
naming conventions (e.g., gsmith – first letter of first name followed by
last name, smithg – last name followed by first initial, etc.). Other
accounts such as the guest account should be disabled and substituted
with a less obvious account name. Other default accounts that should be
disabled and substituted with new site specific accounts are field maintenance, testing, and database administration.
✔
INFOSEC Best Practice #85
Do not permit users to log on locally to a server.
If users other than an administrator have physical access to a server,
then there are a number of things that they can do to try to break into the
machine. Keep servers physically secure and do not permit local logon
of users other than the system administrator. Some systems differentiate
a local account database and network account database. Users should be
authenticated onto the network using a network accounts database, not a
local accounts database.
✔
INFOSEC Best Practice #86
Use a password to access BIOS setup.
79
Information Security Best Practices – 205 Basic Rules
To prevent users from getting around some of your security by booting an operating system from floppy disk, disable booting from floppy in
BIOS and password protect your BIOS if your computer allows.
✔
INFOSEC Best Practice #87
Use a screen saver with a password.
When stepping away from your desk for a short period of time use a
screen saver that kicks in after a few minutes of idle time and locks your
computer with a password. This will prevent users from browsing your
computer when it is unattended.
✔
INFOSEC Best Practice #88
Log off the computer when you are finished using it.
In order for your operating system security to have effect, log off
your computer when you are done using it at the end of the day or if you
will be away from your desk for a longer period of time.
✔
INFOSEC Best Practice #89
Do not keep copies of passwords for system access or
decryption on the same machine that uses the password.
80
Operating System Security Rules
It is a good practice not to keep password lists on computers, but if it
is necessary, then the password(s) must not be kept on the same machine that uses the password for access or decryption. Hackers look for
files containing passwords that can gain them access to other machines,
so don’t make it easy for them.
9.4 FILE SYSTEM PROTECTION
File system protection is the next logical step in securing an operating
system for use by multiple users. This is considered to be one of the key
security requirements for C2 security. A poorly protected file system
allows hackers or unscrupulous employees to gain access to files and
data that may contain sensitive information.
✔
INFOSEC Best Practice #90
System administrators must have full access to all files.
Since system administrators are called upon to help users with a
variety of problems, they must be able to get at user files. This makes it
important that system administrators be chosen carefully for their trustworthiness. Often system administrators are chosen strictly for their
academic and experience qualifications, but they should also be evaluated on their background in terms of trustworthiness. The system
administrator can be the biggest security threat to the organization if not
chosen wisely. Perform background checks on system administrators.
81
Information Security Best Practices – 205 Basic Rules
✔
INFOSEC Best Practice #91
Limit remote server administration to the system
administrator.
Remote administration of servers is becoming a necessity in many
organizations given the proliferation of servers and the reduction in
information systems staff. Often, system administrators are required to
administer more than one system for many departments within the
organization. Remote administration allows the system manager to
manage remote systems that are spread out throughout the organization
without being physically present. Remote administration helps the
system administrator be more efficient at the job and to solve problems
quicker. Also, if there are local pseudo-system administrators that take
care of the day-to-day tasks, then they may inadvertently have access to
more machines within a domain than desirable. Often for convenience
reasons, security is relaxed and the system administrator divulges the
password for remote administration of some of the machines either by
pressure from departments or just for convenience. Under this pressure
the system administrator may divulge the password to solve a near-term
problem, but in turn opens up a long-term security hole. To maintain
security and not cave into the near-term pressures of any department,
you must refer back to the official corporate security policy approved by
the administration.
✔
INFOSEC Best Practice #92
Perform server administration at the console.
82
Operating System Security Rules
If system administration activities are taking place at the system
administrator’s office, then people can walk into the office and potentially see passwords, file protections, sensitive data, etc. Also, the
system administrator may leave momentarily without securing the
computer and thereby allow privileged access to the system. Therefore,
system administration activities should preferably be done in a restricted area.
✔
INFOSEC Best Practice #93
Set file access rights for groups of users.
All files must be accessed using group access rights. Individual
users must belong to groups. Groups can be organized in a hierarchical
group structure. Files must not have rights associated for individual
users since it is difficult to track, clean-up and change files distributed
throughout the system. If a user should not access a file, then that user’s
account should be transferred to a more restrictive group. Cleaning up
file protections becomes much easier.
✔
INFOSEC Best Practice #94
Restrict network file access by network share permissions
that are granted on a need to use basis only.
Files accessed via a network connection on virtual drives must have
network file access permissions. This is another layer of security on top
of file protection. Network access of files on a virtual drive must be
83
Information Security Best Practices – 205 Basic Rules
group controlled. Both network share permissions and individual file
permissions are used to restrict access to a file. This multi-level access
control is very important because there are many users on an
organization’s entire network, including the internet, that may have
access to virtual file services. Virtual file services on single user operating systems such as MS Windows 98 should be protected with
passwords. An example of virtual file services are NFS (Network File
System) a UNIX based system which has also been ported to Microsoft
NT and 95, and Microsoft NT’s network shares using SMB (System
Message Block) communication.
✔
INFOSEC Best Practice #95
Set default file protection to be as restrictive as possible.
When installing a new operating system, the system administrator
must first protect the entire file system for his own access only, and then
incrementally open directories and files to those users with a need to
access. Some operating systems have default protections open to everyone. The system administrator should not rely on defaults, but secure
the system according to his or her own file protection best practices. Set
file permissions for system administrators to full access (i.e., Read,
Write, Execute, Delete, Change Ownership, etc.), but selectively restrict
access for users based on their type of access needs. If a user only needs
to execute a file, then give the group that the user belongs to “Execute”
permission only. If a user should have the option of editing a file, provide “Write” access. Any combination of permissions may be given, but
should be as restrictive as possible.
84
Operating System Security Rules
✔
INFOSEC Best Practice #96
Prevent users from viewing all directory names down a
directory tree.
All directory names in a directory tree should not be seen by those
users that do not have a need to access files at that directory level. The
user should not have the option of exploring directories throughout the
system in order to get clues to the type of information that is stored
within those directories. Therefore, set permissions on directories so
that users can have access down a directory tree without seeing the
name of unauthorized directories. For instance, the higher up the directory hierarchy a user goes, the closer the user is to system related
directories.
✔
INFOSEC Best Practice #97
Use access control lists to restrict access to files by
individual users.
To further restrict access to files by a specific user, use access control
lists to specify the type of access a user can have for a particular file.
Even though a user may belong to a group that has full access to a file, an
access control list can restrict a user specifically. When a user attempts to
access a file, the operating system checks the user’s access tokens to
determine if any further restrictions found in the list should apply.
85
Information Security Best Practices – 205 Basic Rules
✔
INFOSEC Best Practice #98
Clean the swap or page file before releasing a computer to
another party.
If a computer is reassigned to another user and it had originally
stored sensitive data, then clean the swap or page file on the system.
The swap or page file may still contain sensitive data that was paged out
of physical memory to the pagefile on disk. There are software programs available that clean these swap or page files.
✔
INFOSEC Best Practice #99
Do not install more than one operating system on your
computer.
If a less secure operating system is installed on your computer, then
access can be granted by booting the less secure operating system. Also,
do not have multiple versions of the same operating system for the same
reason. Usually, the most current operating system is the most secure.
This does not apply for developers working on projects requiring multiple operating systems.
9.5 VIRUS PROTECTION
✔
INFOSEC Best Practice #100
Use a virus scanner on every computer.
86
Operating System Security Rules
There are approximately 100 new viruses created every month. If
you exchange files with other users or are connected to the internet,
then you are susceptible to viruses. Even software officially distributed
by vendors has been known to contain viruses. Since viruses can immobilize your computer and your productivity, destroy files on your
system, and even damage the hardware on your system; it is imperative
that virus protection be deployed on every computer. Viruses often
create unusual, hard to duplicate symptoms that are different to diagnose. The user often does NOT guess that the cause may be a virus
because the symptoms appear as if there may be a bug in the program or
that the user is doing something wrong. Countless hours are lost trying
to fix the problem or work around it. The best protection is to immunize
the system with a virus scanner.
Each server must have virus-scanning software that scans any files
from external sources. Anti-virus software must scan all incoming IP
traffic from the internet, including incoming email and have support to
scan some of the more common word processing files and macros. Also,
if the server is used to backup any remote machines, the backup software should have a built in anti-virus scanner to scan files before being
backed up. A backup report should note if viruses were found.
✔
INFOSEC Best Practice #101
Perform virus scanning of all IP packets at the bastion
hosts and the gateway.
87
Information Security Best Practices – 205 Basic Rules
Scan all IP traffic coming into the site for viruses before it is sent to
any machine on the internal network. Scanners should pick up most
viruses if their databases are maintained on a monthly basis. Macro
viruses embedded in word processing or spreadsheets must have virus
scanners that can detect these types of viruses.
✔
INFOSEC Best Practice #102
Download the latest virus signatures for your virus scanner
software every two weeks.
Since there are so many new viruses that are created each month,
anti-virus software may become obsolete in a relatively short period of
time. When you purchase anti-virus software you are usually entitled to
download copies of the newest virus databases from the vendor’s
internet site. The vendors usually update their virus databases a couple
of times per month. It is recommended that the user copy the software
from the vendor’s internet site to make the anti-virus software current
and effective against the newest viruses. Most virus scanners have the
capability of automatically downloading updates and virus signatures on
a set schedule. System administrators and users often forget this task,
therefore an automated download method should be used. Alternatively,
it is recommended that the system administrator maintain the latest
virus signatures on a server and that users have their PCs configured to
directly access these files. The system administrator can also push these
files down to user PCs. Do not get secondhand copies of the anti-virus
databases; use the databases from the vendor directly.
88
Operating System Security Rules
✔
INFOSEC Best Practice #103
Perform virus scanning of each machine on a designated
schedule.
Install the anti-virus software so that it starts when the system boots
and is resident in memory when the system is running. This will provide
continuous monitoring for viruses that may arrive at the computer either
via floppy disks, mail, or the internet. A virus scanner will not guarantee a virus-free computer, but will significantly reduce the chance of a
virus being present. Occasionally, some programs are incompatible with
the anti-virus software. If this is the case, you must run your anti-virus
software manually on a regular schedule once a week. Servers must be
running anti-virus software continuously.
9.6 NETWORK FILE SHARING SECURITY
Most network file systems have some type of security weakness. The
administrator must be aware of these problems and decide between the
convenience of having network file services and the security holes that
may be open when these services are used.
✔
INFOSEC Best Practice #104
Do not have virtual file services enabled for bastion host
machines.
89
Information Security Best Practices – 205 Basic Rules
Virtual file services can be a method of accessing multiple machines
once a single machine is compromised. Especially, do not have file
services across your firewall. The firewall will need to have a port(s)
open for that service and thereby provide a hole for entry into your
internal network.
✔
INFOSEC Best Practice #105
Deploy NFS on an internal network, separated from a
public network via a firewall.
If you need to use NFS (Network File System), then it must be
deployed on an internal network because it uses simple clear-text authentication (e.g., host name user ID and group ID). This may be barely
acceptable in secure networks where there is limited access to the network, but is not secure on public networks. UNIX systems will typically
use this network file system to share files between users and to set up
virtual drives for users. If NFS is to be used in an organization, then it
should at a minimum be deployed behind a firewall separating the
internal network from an external, public network. To increase security
when using NFS, but with a performance decrease, use Secure NFS
which uses Secure RPC and AUTH_DES encryption.
✔
INFOSEC Best Practice #106
Do not use X-Windows for mission-critical applications
over a public network.
90
Operating System Security Rules
X-Windows cookies are transmitted over the network in clear text.
Therefore, applications with sensitive data must not use X-Windows for
network-based windowing on a public network.
9.7 NETWORK SOFTWARE
✔
INFOSEC Best Practice #107
Strictly control IP assignment.
One of the key pieces of information for a hacker is knowledge of
the IP address ranges within an organization. Additionally, knowledge of
the exact IP address of servers and other mission-critical machines can
help a hacker focus an attack on these systems. IP addresses can be obtained by buying a block of addresses from Internic or from an
organization’s Internet Service Provider. Addresses can be bought as class
C licenses (256 addresses) or class B licenses (32,000 addresses). Class B
licenses are usually reserved for service providers and large companies.
There are very few class B licenses around and they are very expensive.
Almost all companies will purchase Class C licenses as blocks of 256. If
there is no connection to a public network using IP and there are no plans
to hook into such a public network in the future (highly unlikely), then the
company can use their own IP addresses when setting up systems. Also, if
firewalls and proxy servers are used to separate the internal network from
the public network, then all the systems internal to the company do not
need legal IP addresses since the proxy server will use its own address to
communicate with the public network. Most companies buy blocks of
legal licenses because they do not know how they might want to commu-
91
Information Security Best Practices – 205 Basic Rules
nicate in the future and therefore do not want to have to reconfigure their
computers with new IP addresses. You must have legal IPs if you have
any machine connected to a public network. Home PCs and small offices
will get their IPs through their service provider.
✔
INFOSEC Best Practice #108
Use host to host authentication between two logical
networks.
Operating systems currently do not use security protocols to establish
host to host links to mail servers or file servers. By not having authentication software that can determine whether the host you are connecting to is
really who it claims to be, the user is vulnerable to an attacker that is
mimicking the destination host. Kerberos V5 should be used between two
hosts that need to communicate for server based applications. Kerberos 5
is a server-based authentication mechanism that: 1) receives a message
containing a user’s username and current time encrypted with the user’s
password, 2) looks up the username to determine the password and, 3)
tries to decrypt the encrypted time that was sent. If the Kerberos server
can decrypt the current time, then it creates a ticket granting ticket containing a session key and ticket for the Kerberos Ticket Granting Service,
encrypts it with your password and sends it to you. In this way the user’s
password is never transmitted across the network and public key cryptography is not used.
92
Operating System Security Rules
✔
INFOSEC Best Practice #109
Start up processes only for desired TCP applications.
Each process (daemon) that is running, but is not used is a potential
vulnerability to the system. TCP/IP services such as FTP, Telnet, and
others use TCP ports to communicate over the IP network. Potential
hackers can try to access these ports and exploit the vulnerabilities of
these TCP programs. Also, custom applications and database applications use specific TCP ports for client-server applications. These custom
written programs may be more vulnerable than hardened TCP applications and provide a potential point for break-in to the machine. If these
programs are running from an administrative account, then the hacker
may have full access to the machine. To reduce the risk of potential
break-ins through TCP ports, only run those TCP applications that you
must use on the machine.
9.8 SECURITY LOGS
✔
INFOSEC Best Practice #110
Keep a security event log on your computer.
A security logger should be available as part of the operating system
or as a third party product. All servers and workstations with restricted
access must run a security logger. The security log must be turned on
and record the following categories of events:
93
Information Security Best Practices – 205 Basic Rules
1. Logins and logoffs.
2. Object access (logging this data may lead to very large log
files).
3. File permission changes.
4. File ownership changes.
5. Security policy changes.
6. Changes in user rights.
7. Group changes for an account.
8. System restarts and shutdowns.
9. Virus scans.
Security logs can grow to take up much disk space therefore the
server must have a large enough disk. Periodic maintenance will need to
be performed to archive or discard the logs.
✔
INFOSEC Best Practice #111
Review the security log on each server on a daily schedule.
Recording security events is a waste of resources if the system
administrator does not review the logs on a daily basis. All it takes is
five minutes each day to invoke the security log viewer and scan
through the recorded events. Since most machines will have little
hacker or mischievous activity, system administrators get complacent
and stop reviewing the security log. If the machine is important enough
for you to decide that security logging must be turned on in order to
94
Operating System Security Rules
comply with your established security policy, then the administrator
must treat this task seriously. To ease the tedious nature of this task,
there are security monitors available for some operating systems that
review the security log automatically according to your designated
criteria and issue an alarm about potential security breaches. There is a
cost involved in purchasing such software, but it may help in enforcing
the security policy at your organization.
✔
INFOSEC Best Practice #112
Archive all security logs.
Security logs should be archived on a periodic basis (i.e., yearly).
Keeping archives will allow you to go back and check for suspicious
activity (e.g., file and object accesses) that occurs on the system over
time.
✔
INFOSEC Best Practice #113
Record security events on your firewall.
Your firewall computers must have security logging turned on.
Firewall software has its own security and network event logging which
must be turned on in addition to operating system security logging.
95
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement