Excerpts

Network engineers, managers, programmers, professors and students have all asked how they can gain a deeper understanding of computer networks and internets. This book answers the question. It asserts that the best way to learn is by doing--there is no good substitute for hands-on experience with a real network. Interconnecting hardware, configuring network systems, measuring performance, observing protocols in action, and creating client-server programs that communicate over a network all help sharpen one's understanding and appreciation. What hardware and software facilities are required for hands-on experimentation? Instead of specifying an exact platform, this book is organized into six sections that each consider a hardware platform and outline experiments that can be carried out using the hardware. The first section begins by considering the smallest possible facility, a single stand-alone computer. Successive sections describe increasingly more powerful (and more expensive) facilities and the experiments they support. The last sections document advanced hardware and software facilities used for protocol development and network systems engineering. The point is that experimentation is always possible--although the facilities at hand determine the types of experiments that can be performed, even low-cost, general-purpose facilities offer opportunities. The broadest distinction among facilities concerns isolation. Early sections of the book describe experiments that can be carried out on conventional, general-purpose computers connected to a production network. Later sections describe experiments such as packet capture, intranet configuration, and protocol development that require a special, dedicated facility. Industry often uses the terms testbed or testnet to describe a separate dedicated facility; academia usually uses the term laboratory. An industrial testbed can serve two purposes. Like an academic laboratory, a testbed provides an environment that supports training. In addition the testbed provides a safe environment in which new or upgraded network systems can be configured, measured, and tested before being installed in the company's production network. Although we use the academic term laboratory throughout the book, many of the experiments are designed with industrial testbeds in mind. In the section on configuration, for example, experiments specify using an isolated facility to configure hosts, routers, and a firewall to form an intranet. In addition to a wide variety of topics, the experiments in this book cover a wide range of difficulty. Some experiments, especially those near the beginning of each chapter, are straightforward and may require less than a half hour to perform. Other experiments are both difficult and lengthy. For example, the IP router experiment described in Chapter 20 is taken from a second-year graduate class that I teach at Purdue. Students work in teams and require most of a semester to build a working IP router. Most experiments list optional extensions that suggest ways to go beyond the basics. The best students in my classes work through all the options, and sometimes invent options of their own. Networking professionals who are working alone can pick and choose among experiments in various chapters. Programmers will focus on the client-server experiments in Parts I and II; system administrators will focus on the configuration experiments in Part IV. Engineers who implement and optimize network systems and protocol stacks will focus on the performance measurements in Part III or protocol development experiments in Parts V and VI. Professors teaching networking courses can choose experiments appropriate to the class. Most colleges and universities cram all of networking and internetworking into a single one-semester overview course. In such a course, students should see a wide range of experiments to acquaint them with all aspects of the subject, especially network programming, network measurement, and protocol analysis. For example, students in my undergraduate class begin the first week using Internet applications and writing programs as described in Part II. When the lectures cover Local Area Networks, the students are assigned measurement experiments from Part III. The weeks that the lectures cover IP and TCP, students perform packet capture and protocol analysis experiments from Part III. Finally, students are ready to tackle a more advanced (socket) programming experiment from Part III. Colleges and universities fortunate enough to have multiple courses in networking can divide the experiments in this book among the courses and go into more depth in each. In a one-semester network programming course students concentrate on the programming experiments in Part III. Students in such a course also enjoy building the internet emulation gateway from Chapter 8 and constructing a library for the simplified API. If students do not have any previous experience, they can begin by using the simplified API in Part II, which allows them to start programming before they learn the details of sockets. I usually encourage students in a network programming course to incorporate the optional extensions and to think about ways the software can be parameterized. In a one-semester network administration course, students can focus on the configuration experiments in Part IV. Students enjoy creating their own version of an intranet, and are especially pleased when they can configure unconventional domain names. As an interesting challenge, I ask students to establish two apparently unrelated domain hierarchies for the same set of computers. Finally, in any course that discusses protocol design, students should begin by building the Internet emulation gateway described in Chapter 8. The gateway project requires students to extend their knowledge of socket programming to include UDP, allows them to see an application gateway, and causes them to consider the possible errors with which protocols must contend (i.e., packet loss, duplication, corruption, and delay). Excerpted from Hands-On Networking with Internet Technologies by Douglas E. Comer All rights reserved by the original copyright owners. Excerpts are provided for display purposes only and may not be reproduced, reprinted or distributed without the written permission of the publisher.