Security in Distributed Network

M.S. Nisha Ms. Rashmi Saboo

JIET, NH-65, Pali road, Near Mogra, Jodhpur- 342008
(ms.nisha@jietjodhpur.com, seemsab@gmail.com )

access network resources. For other

Abstract NOSs the user logs into a server. In the
This paper tries to figure out the various case of the latter, the server is the
mechanisms that can be applied for the network. Certainly the discussion
security of the distributed network. It covering guidelines for passwords can
also lists the various means by which the be applied directly to network
biometrics can be used for authentication authentication.
and providing security in a network.
2.2 Controlling End User Access
1. Introduction Creating an account and assigning a
Security for stand-alone password are only small parts of giving
computers and small networks was someone access to the network. A
handled by physical security and by network administrator also has to
logging into computers and domains. determine other account parameters such
With open networks like the Internet and as when an end user can access the
pervasive environments, issues network, what groups the user is
concerning security and trust become associated with, what files he or she can
crucial. There is no longer the physical access, and limitations on network and
aspect of security due to the distributed server resources.
nature of the networks and the concept
of user authentication to a domain is not 2.3 Concurrent Logins
possible. In such an open, Users should not be allowed concurrent
heterogeneous, distributed environment sign-on privileges. Once an end user has
there is a great likelihood that logged into a network they should not be
inconsistent interpretations will be made able to login somewhere else without
of the security information in different first logging out from where they
domains. Existing security originally logged in. The only exception
infrastructures deal with authentication to this rule should be the LAN
and access control. These mechanisms administrator and his or her backup.
are inadequate for the increased While anybody recognizes that this
flexibility required by distributed could cause operational problems for
networks. some users, there are several reasons for
limiting concurrent sign-ons. First, it
2. Security Modes saves network resources, such as
2.1 LAN Guidelines memory and licenses. It can also prevent
It is often difficult to distinguish where the unauthorized use of an account, so
the individual server ends and the long as the user is logged in. It also
network begins. Some Network prevents the user from forgetting to log
Operating Systems can be configured so out. When we allow concurrent sign-ons,
that the end user logs into a domain to the end users often lose track of where
they are logged in and forget to sign off the LAN administrator and his/her
everywhere. Users can leave themselves backup.
logged into the network on a workstation
without even realizing it. They open a 2.5.1 Time/Day Restrictions
window of vulnerability to the network Employees who are authorized to access
and themselves when they leave and use sensitive and/or confidential
accounts signed on. One solution to this data can access it during office hours
problem is to implement a process that only. If an employee does not normally
automatically logs off inactive users. work in the evenings and on the
There are also systems that freeze a weekends, then the ability to access the
workstation or lock a keyboard on an network should be restricted for that
inactive session, after a specified period time period. Most every operating
of time. To release the keyboard lock the system and NOS has the capability to
user must enter a password. restrict an account's access to specific
time periods.
2.4 Available Disk Space
It is important to limit the amount of 2.5.2 Access to Directories and
disk space allocated to each end user. Trustee Rights
giving users unlimited disk space may Users should only be given access rights
end up requiring the purchase of to directories they need to function. If a
additional disk capacity. We have seen user needs temporary access to a
situations where users crashed servers directory, the access rights should be
because their accounts did not restrict removed when they are no longer
the amount of disk space the user was needed.
allowed. In one instance, a user was Users should only be given the trustee
running a report that spooled a massive rights they need to do their job. Once a
file to disk. The result was that all the right is no longer required, remove it
available space was consumed, and the right away. Trustee rights should be
server crashed. Users should also be audited periodically.
encouraged to clean up their directories
on a regular basis. We recognize that the 2.6 File Attributes
comparative cost for disk drives File-access attributes, such as read,
continues to drop to where the cost per write, execute, and delete, should be
megabyte is nominal, but disk drives still granted based on need. In addition, files
need to be backed up. containing confidential or sensitive
information should be restricted to a
2.5 Restrictions to Location or minimum number of users. File
Workstation attributes for executables should be
Consideration should be given to restricted. End users should only have
restricting, to a specific workstation, end read access to those files that are needed
users who are authorized to enter to function. Particular attention should
sensitive transactions or who perform be paid to operating system executables.
particularly sensitive and/or confidential If file attributes for executable files are
work. It is preferable to locate the station not restricted, the executable files can be
in a restricted area. Obviously, access to modified. With loosely defined file
the server itself should be restricted to attributes, important executable files can
be changed or replaced with Trojan 2.9 Single Sign-On
horse programs.
Presently, every morning we enter in
2.7 Other Privileges multiple usernames and passwords to
Network commands and executables gain access to the various networks,
should be restricted to administrators, systems, and applications that we need
auditors, and security personnel. With in order to perform my job. we have a
certain operating systems, such as password for the NT domain, a password
Windows NT, consider renaming the for the Novell server, different
administrator account to something else. passwords for different UNIX systems, a
That way a potential hacker won't know password for my e-mail, and passwords
the name of the privileged account. for various applications.
One of the things that every system Having so many passwords can be
administrator fears most is a hacker confusing to end users and can actually
gaining privileged administrative access create vulnerabilities, because the only
to a system over the network. Privileged way the end user can remember the
accounts such as administrator for NT or passwords is to write them down. One
root for UNIX should not be allowed to alternative to using multiple passwords
login over the network. Network access is the use of a single sign-on (SSO).
to the administrative account can be With an SSO system users are only
restricted in different ways for different required to authenticate themselves
operating systems. once. Once users have authenticated
themselves the SSO system handles the
2.8 Remove Inactive Accounts management and access to other network
Organizations should review network resources, such as servers, files, and
user accounts on a regular bases and applications.
delete any accounts that are no longer
required. Accounts for users or While policy-based management can
employees no longer with the have advantages it also holds risks.
organization should be deleted. Firms When Windows 2000 was first released
should also delete inactive accounts, there was much debate about the security
removing or disabling username of Active Directory. When utilizing
accounts that have not been accessed in Active Directory Services, there are
the last three to six months. Hackers dangers associated with loosely defined
frequently try to exploit inactive policies or the granting of broad
accounts for the initial break into a administrative privileges to managers
system or as a means to gain access to a and administrators, which can result in
network again. They know they can alter gaping holes in an organization's
an inactive account, by changing the network security. At the very least it can
password, for example, without fear of result in potential exposure of
the change being detected by the user of confidential information. Due to the
the account. The guest account should design of Active Directory,
not have the same privileges as normal administrators who have been restricted
accounts. from accessing particular network
objects can actually take ownership of
the restricted objects with a few clicks of
a mouse. Microsoft's response to the
flaw was to recommend that multiple
domains be implemented with Active 2.11 Static IP Addresses Versus
Directory, which defeats the purpose of Dynamic Host Configuration Protocol
implementing the directory service. It is (DHCP)
amusing to note that Microsoft's initial DHCP enables network administrators to
response to the flaw was to call it a centrally manage and automate the
"feature" of Active Directory. We guess assignment of IP addresses for an
it's true that any software bug organization's network. This means that
sufficiently advanced becomes a feature. a computer with a DHCP client can
However, Active Directory is not alone dynamically obtain an IP address from a
in containing risks associated with remote server (DHCP server). Each time
loosely defined policies. The same a workstation logs into the network it is
danger is associated with any policy assigned an IP address. An alternative
based system and can result from poorly would be to use preassigned static IP
defined or implemented policies. With addresses that each system would be
Active Directory, the risk is heightened individually configured to use. The
by the concern that organizations will major advantages of DHCP include
attempt to implement it with the same simplicity of configuration for the
broad privileges that they had clients, more efficient assignment of IP
implemented NT domains. Active addresses, and ease of administration.
Directory and NT domains are two The major disadvantage of DHCP is that
entirely different systems with different the assignment of IP addresses is
approaches to security and implementing temporary. There are other alternatives
them in the same manner can have to DHCP, such as Reverse Address
disastrous results. Resolution Protocol (RARP) or
Bootstrap Protocol (BOOTP), that
2.10 Honeypot Systems essentially function the same way.
One technique that many administrators
employ is the use of "honeypot" 2.3 Biometrics
systems. They are basically deception 2.3.1Identification and Authentication
systems that contain phony services,
files, and applications designed to There are three basic methods employed
emulate wellknown holes with the goal for identification and authentication.
of entrapping hackers. They are designed They are listed as follows:
to attract hackers, hence the name • Something you know
"honeypot." • Something you have
The honeypot is intended to make • Something you are
hackers believe that they have
discovered a real system. The system is 2.3.2 Biometric Identification and
designed to lure a hacker into a "safe" Authentication
network or server that impersonates
important applications or information. Biometric authentication is the process
When the hacker enters the honeypot the of using some physical characteristic,
trap is sprung and the alarm is sounded. trait, aspect of physical being, or
behavior to authenticate one's identity.
The most commonly known example is An effective biometric system needs to
the process of employing fingerprints to be able to handle temporary
identify an individual. Biometric physiological changes. The backup
authentication usually fits into one of methods to authenticate users in the
two general categories. The first is event the biometrics fail.
physical characteristic recognition
(PCR), which relies upon a physical 2.3.5 Environmental Conditions
characteristic such as a fingerprint, retina The environment in which it will operate
or iris scan, voiceprint, or facial will effect a lot. Water, noise, moisture,
geometry for identification and and dirt can adversely impact the
authentication. The second category is operation of some biometric
behavioral characteristic recognition authentication systems. A factory floor
(BCR). BCR relies on behavioral where workers get their hands dirty with
characteristics such as how a person grease or where the conditions are very
types at a keyboard, writes, or signs his wet would not be the best environment
or her name. to install a fingerprint reader or hand
reader. A fingerprint reader or hand
2.3.3 Biometric Identification scanner would be equally ineffectual in
Reliability an environment where workers wear
gloves.
There are two critical characteristics that A retina scanner or face geometry reader
reviewed before deploying any system. would not be advised in an environment
They are listed as follows: in which individuals must wear
protective eyewear or masks. In
• False acceptance rate (FAR) addition, voiceprint readers would not
• False rejection rate (FRR) work well in noisy environments.

The FAR is the rate at which a system 2.3.6 User Acceptance

incorrectly accepts or recognizes a To achieve a successful deployment of a
would-be user as authorized to access biometric authentication system, it is
the system when in fact he or she are important to gauge user acceptance of
not. the technology being used. Users may be
uncomfortable with retina scanners and
The FRR, the rate at which a system find recording of fingerprints an
incorrectly rejects a legitimate user. The invasion of privacy. Another issue that
FRR is important to the successful needs to be considered before deploying
deployment of any biometric a system is general hygiene.
authentication system. If the FRR of a
system is too high, it can cause end-user 2.3.7 Security of the Biometric System
frustration. The frustration can lead users Another critical factor with biometric
to circumvent proper authentication identification and authentication systems
procedures to avoid the biometric is how it handles communication and
system. storage. The mathematical representation
of fingerprint is sent to the server for
2.3.4 Backup Authentication identification and authentication.
Another method would be to store the security”
template on the local system. In addition, 4. John D. Woodward (Jr.), Nicholas M.
storing the template on a local hard drive Orlans, Peter T. Higgins “ Biometrics”
would introduce other security issues. 5. Anil K. Jain, Ruud Bolle, Sharath
If the templates are stored on a server we Pankanti “Biometrics: personal
also need to consider how they are identification in
stored and the security employed to networked society”
prevent them from being compromised.
These are the types of issues that an
administrator needs to take into account
before deploying any biometric
identification and authentication system.

2.3.8 Interoperability
Another issue that is much more difficult
to resolve and we believe will be around
for a while is the fact that there is no
interoperability between biometric
systems. Every single product on the
market is proprietary. It is also difficult
to find a product that has operating
system interoperability. As a result, if we
work for a large organization, we will be
hard pressed to find a system that can be
deployed across the entire enterprise.

2.4 Conclusion
Security is the need of current world.
Our paper discusses the different
methods which can be used to ensure
security in network. Security is
everybody's business, and only with
everyone's cooperation, an intelligent
policy, and consistent practices, will it
be achievable.

2.5 References

1. Joseph Migga kizza “Computer

network security”
2. Secure Computers and Networks
Analysis, design and implementation
“Eric A. Fisc, Gregory B White”
3. Gregory B. White, Eric A. Fisch, Udo
W. Pooch “Computer system and
network