Security in the Local Area Network
In the local network, you are in charge of which protocols are used and you can set up monitoring and auditing so that you can track abuse. Even in a large company network, tools based on the Simple Network Management Protocol (SNMP) or RMON provide centralized control and security tracking.
What Is RMON?
RMON stands for Remote Monitoring. RMON is a data gathering and analysis tool that provides for remote monitoring of network components, such as hosts, routers, switches, and other network devices. The functionality of RMON, which is described in RFCs 1513 and 1757, extends the monitoring capabilities that SNMP provides.
The most basic form of security used in a local area network is the user account and password. When implemented in a secure manner, each user who accesses the network has a unique account name (or username) that is assigned by the network administrator and a password that is usually under the control of the user himself. This form of authentication has its roots in early standalone (non-networked) computers that ran multiuser operating systems. The unique username allowed the operating system to keep track of which user was actually logged in to the computer, which made possible the ability to grant differing rights and privileges to users depending on their needs. An initial password is created by the administrator who creates the user account. It is the responsibility of the user, however, to change that initial password to a value that is difficult to guess. It is also the responsibility of the user to safeguard the password so that it does not become known to anyone else.
Using a Shadow Password File
One way to enhance security of the passwords usually stored in the /etc/ passwd file is to use a shadow password file that is not world-readable. Check the documentation for your UNIX distribution to see whether this feature is available.
When this sort of authentication is incorporated into local area networking technologies, it can become a little more complicated. When only one computer is involved, it is very easy to maintain a single secured database that stores usernames and passwords that the operating system can use to check against users. When multiple computers are linked together in a network, where does this security information reside?
In some operating systems, such as UNIX, each computer in the network has its own password databasethe file /etc/passwd. This file stores usernames and passwords, along with other information pertinent to the user account. Although this file is a simple text file that can be printed out and examined, it is still relatively secure because the passwords stored in the file are encrypted. That is, it is secure if you trust the users on your network. Programs such as Crack can be used to discover what these encrypted passwords are, so as you can probably guess, this method of ensuring secure authentication leaves something to be desired when connecting to the Internet.
Synchronizing Password Files
While UNIX has many built-in networking components, one thing that it lacks is a method for synchronizing password files between multiple systems on a network. In Windows NT domains, one or more servers act as domain controllers and process user logons for the domain. Under UNIX, each server has its own password file. If users are to log in to multiple systems on the network, they must have an account on each system. To fix this shortcoming, an application called Yellow Pages was developed by Sun to keep important system files synchronized. Because of a trademark dispute, Yellow Pages was renamed NIS, for Network Information Service.
It should become obvious at the start that this method of using a different security database for each computer in the network has its drawbacks. For example, a good security policy dictates that the user must change her password after it has been in use for some time. If the user needs to access resources on multiple computers in the LAN, the user must change the password on more than one computer when it comes time to make a change.
Network operating systems can provide a solution by using a login database that contains accounts for all users on the network. In Windows NT, the concept of a domain is used to group users into a central Security Accounts Manager (SAM) database. When a user logs on to a Windows NT computer that participates in the domain, a dialog between that user's workstation and a server operating as a domain controller authenticates the user. Early version of Novell NetWare enabled users to log on to each server they accessed. Novell Directory Services (NDS) now provides a centralized user database.
When a connection is made to the Internet, the problem of user authentication becomes more problematic. When mobile users are away from the office, should they use a modem to contact the home office, or can the Internet be used? If the Internet is used, how can you ensure a secure login session? There are many solutions to this problem, from Kerberos and public key encryption technology to virtual private networks (VPNs).
User authentication is the first thing most of us think about when it comes to computer security. After a user has been authenticated to the computer or the network, the next mechanism that is used to control access is resource protections. UNIX, Windows NT, Novell NetWare, other operating systems, and network operating systems all provide some form of resource protections that enable you to specify which files, directories, printers, and so on, are accessible to users. Resource protections can be used to restrict which users can access objects, and auditing techniques can be used to track usage as well as abuse of privileges.
Like the user authentication databases, the systems and network administrators of the LAN are in control of resources on the systems connected to the network. This control, however, is based on the assumption that your user authentication methods are totally secure and that you have configured resource protections correctly.
I now know that simple authentication methods are easy to break, given enough time. Methods that were sufficient in a small LAN are not sufficient when you connect to a much larger network. If authentication methods fail, resource protections mean very little because an attacker can pretty much become whichever user he wants.
Resource protections do not provide complete security against intruders. With an Internet connection, you expose your network to many more potential intruders than you can imagine. To protect network resources, further measures must be taken. Firewalls, the topic of this book, are the next major step.
When dealing with computers in a local area network, it is relatively easy to make sure that the physical site is secure. Using a computer room with a locked door for important servers can usually prevent them from being compromised. When the LAN spans multiple buildings, as in a college campus, physical security is still under control of someone locally.
When a network is connected to the Internet, it becomes part of a much larger network over which you have no physical control. You might have a little influence with your ISP and might be able to control how the router is configured for your service, but you have no control over the millions of other Internet host computers, routers, switches, and so on. The firewall is the method you use to create a strong front door that, hopefully, will keep out those who want to harm the network.