Stepping Through..

One of the limiting factors in the operation of large Novell Internetwork Packet Exchange (IPX) internetworks is the amount of bandwidth consumed by the large, periodic Service Advertisement Protocol (SAP) updates. Novell servers periodically send clients information about the services they provide by broadcasting this information onto their connected local-area network (LAN) or wide-area network (WAN) interfaces. Routers are required to propagate SAP updates through an IPX network so that all clients can see the service messages. It is possible to reduce SAP traffic on Novell IPX networks by the following means:

•Filtering SAP updates through access lists. SAP updates can be filtered by prohibiting routers from advertising services from specified Novell servers.

•Configuring Cisco routers on Novell IPX networks to run Enhanced IGRP. Although filters provide a means of eliminating the advertisements of specified services, Enhanced IGRP provides incremental SAP updates for a finer granularity of control. Complete SAP updates are sent periodically on each interface only until an IPX Enhanced IGRP neighbor is found. Thereafter, SAP updates are sent only when there are changes to the SAP table. In this way, bandwidth is conserved, and the advertisement of services is reduced without being eliminated.

Incremental SAP updates are automatic on serial interfaces and can be configured on LAN media. Enhanced IGRP also provides partial routing updates and fast convergence for IPX networks. Administrators may choose to run only the partial SAP updates or to run both the reliable SAP protocol and the partial routing update portion of Enhanced IGRP.

•Configuring Cisco routers on Novell IPX networks to send incremental SAP updates. With Software Release 10.0, the incremental SAP updates just described can be configured for Cisco routers on Novell IPX networks, without the requirement of running the routing update feature of Enhanced IGRP (only the partial SAP updates are enabled). This feature is supported on all interface types. Again, SAP updates are sent only when changes occur on a network. Only the changes to SAP tables are sent as updates.

The Enhanced Interior Gateway Routing Protocol (IGRP) combines the ease of use of traditional routing protocols with the fast rerouting capabilities of link-state protocols, providing advanced capabilities for fast convergence and partial updates. When a network topology change occurs, the Diffusing Algorithm (DUAL) used with Enhanced IGRP provides convergence in less than five seconds in most cases. This is equivalent to the convergence achieved by link-state protocols such as Open Shortest Path First (OSPF), Novell Link Services Protocol (NLSP), and Intermediate System-to-Intermediate System (IS-IS). In addition, Enhanced IGRP sends routing update information only when changes occur, and only the changed information is sent to affected routers.

Enhanced IGRP supports three network level protocols: IP, AppleTalk, and Novell Internetwork Packet Exchange (IPX). Each of these has protocol-specific, value-added functionality. IP Enhanced IGRP supports variable-length subnet masks (VLSMs). IPX Novell Enhanced IGRP supports incremental Service Advertisement Protocol (SAP) updates, removes the Routing Information Protocol (RIP) limitation of 15 hop counts, and provides optimal path use. A router running AppleTalk Enhanced IGRP supports partial, bounded routing updates and provides load sharing and optimal path use.

The case study provided here discusses the benefits and considerations involved in integrating Enhanced IGRP into the following types of internetworks:

•IP—The existing IP network is running IGRP

•Novell IPX—The existing IPX network is running RIP and SAP

•AppleTalk—The existing AppleTalk network is running the Routing Table Maintenance Protocol (RTMP)

When integrating Enhanced IGRP into existing networks, plan a phased implementation. Add Enhanced IGRP at the periphery of the network by configuring Enhanced IGRP on a boundary router on the backbone off the core network. Then integrate Enhanced IGRP into the core network.

When most people talk about security, they mean ensuring that users can only perform tasks they are authorized to do, can only obtain information they are authorized to have, and cannot cause damage to the data, applications, or operating environment of a system.

The word security connotes protection against malicious attack by outsiders. Security also involves controlling the effects of errors and equipment failures. Anything that can protect against a deliberate, intelligent, calculated attack will probably prevent random misfortune as well.

Security measures keep people honest in the same way that locks do. This case study provides specific actions you can take to improve the security of your network. Before going into specifics, however, it will help if you understand the following basic concepts that are essential to any security system:

•Know your enemy

This case study refers to attackers or intruders. Consider who might want to circumvent your security measures and identify their motivations. Determine what they might want to do and the damage that they could cause to your network.

Security measures can never make it impossible for a user to perform unauthorized tasks with a computer system. They can only make it harder. The goal is to make sure the network security controls are beyond the attacker’s ability or motivation.

•Count the cost

Security measures almost always reduce convenience, especially for sophisticated users. Security can delay work and create expensive administrative and educational overhead. It can use significant computing resources and require dedicated hardware.

When you design your security measures, understand their costs and weigh those costs against the potential benefits. To do that, you must understand the costs of the measures themselves and the costs and likelihoods of security breaches. If you incur security costs out of proportion to the actual dangers, you have done yourself a disservice.

•Identify your assumptions

Every security system has underlying assumptions. For example, you might assume that your network is not tapped, or that attackers know less than you do, that they are using standard software, or that a locked room is safe. Be sure to examine and justify your assumptions. Any hidden assumption is a potential security hole.

•Control your secrets

Most security is based on secrets. Passwords and encryption keys, for example, are secrets. Too often, though, the secrets are not really all that secret. The most important part of keeping secrets is knowing the areas you need to protect. What knowledge would enable someone to circumvent your system? You should jealously guard that knowledge and assume that everything else is known to your adversaries. The more secrets you have, the harder it will be to keep all of them. Security systems should be designed so that only a limited number of secrets need to be kept.

•Remember human factors

Many security procedures fail because their designers do not consider how users will react to them. For example, because they can be difficult to remember, automatically generated "nonsense" passwords are often found written on the undersides of keyboards. For convenience, a "secure" door that leads to the system’s only tape drive is sometimes propped open. For expediency, unauthorized modems are often connected to a network to avoid onerous dial-in security measures.

If your security measures interfere with essential use of the system, those measures will be resisted and perhaps circumvented. To win compliance, you must make sure that users can get their work done, and you must sell your security measures to users. Users must understand and accept the need for security.

Any user can compromise system security, at least to some degree. Passwords, for instance, can often be found simply by calling legitimate users on the telephone, claiming to be a system administrator, and asking for them. If your users understand security issues, and if they understand the reasons for your security measures, they are far less likely to make an intruder’s life easier.

At a minimum, users should be taught never to release passwords or other secrets over unsecured telephone lines (especially cellular telephones) or electronic mail (email). Users should be wary of questions asked by people who call them on the telephone. Some companies have implemented formalized network security training for their employees; that is, employees are not allowed access to the Internet until they have completed a formal training program.

•Know your weaknesses

Every security system has vulnerabilities. You should understand your system’s weak points and know how they could be exploited. You should also know the areas that present the largest danger and prevent access to them immediately. Understanding the weak points is the first step toward turning them into secure areas.

•Limit the scope of access

You should create appropriate barriers inside your system so that if intruders access one part of the system, they do not automatically have access to the rest of the system. The security of a system is only as good as the weakest security level of any single host in the system.

•Understand your environment

Understanding how your system normally functions, knowing what is expected and what is unexpected, and being familiar with how devices are usually used, help you to detect security problems. Noticing unusual events can help you to catch intruders before they can damage the system. Auditing tools can help you to detect those unusual events.

•Limit your trust

You should know exactly which software you rely on, and your security system should not have to rely upon the assumption that all software is bug-free.

•Remember physical security

Physical access to a computer (or a router) usually gives a sufficiently sophisticated user total control over that computer. Physical access to a network link usually allows a person to tap that link, jam it, or inject traffic into it. It makes no sense to install complicated software security measures when access to the hardware is not controlled.

•Security is pervasive

Almost any change you make in your system may have security effects. This is especially true when new services are created. Administrators, programmers, and users should consider the security implications of every change they make. Understanding the security implications of a change is something that takes practice. It requires lateral thinking and a willingness to explore every way in which a service could potentially be manipulated.

Over the past few years, the concept of end-users being able to send and receive audio and video (known collectively as multimedia) at the desktop has gained considerable attention and acceptance. With high-performance 486, Pentium, and PowerPC CPUs, more than 80 percent of the personal computers sold during 1995 were multimedia capable. Today, it is not uncommon for end-users to run video editing and image processing applications from the desktop.

The proliferation of more and more multimedia-enabled desktop computers has spawned a new class of multimedia applications that operate in networked environments. These network multimedia applications leverage existing network infrastructure to deliver video and audio applications to end users. Most notable are videoconferencing and video server applications. With these applications, video and audio streams are transferred over the network between peers or between clients and servers. There are three types of multimedia applications:

•Unicast—Unicast applications send one copy of each packet to each host that wants to receive the packet. This type of application is easy to implement, but it requires extra bandwidth because the network has to carry the same packet multiple times—even on shared links. Because unicast applications make a copy of each packet, the number of receivers is limited to the number of copies of each packet that can be made by the CPU that runs the unicast application.

•Broadcast—Broadcast applications send each packet to a broadcast address. This type of application is easier to implement than unicast applications, but it can have serious effects on the network. Allowing the broadcast to propagate throughout the network is a significant burden on both the network (in terms of traffic volume) and the hosts connected to the network (in terms of the CPU time that each host that does not want to receive the transmission must spend processing and discarding unwanted broadcast packets). You can configure routers to stop broadcasts at the LAN boundary (a technique that is frequently used to prevent broadcast storms), but this technique limits the receivers according to their physical location.

•Multicast—Multicast applications send each packet to a multicast group address. Hosts that want to receive the packets indicate that they want to be members of the multicast group. This type of application expects that networks with hosts that have joined a multicast group will receive multicast packets. Multicast applications and underlying multicast protocols control multimedia traffic and shield hosts from having to process unnecessary broadcast traffic.

This case study examines multicast protocols that have been developed for the Internet Protocol (IP) and for AppleTalk, as well as Cisco Internetwork Operating System (Cisco IOS) features that can help your network deliver video and audio smoothly.