The Internet is a global communications network. With connections in more than 100 countries, tens of millions of people use the Internet for business, education, and recreation. Electronic commerce is beginning on the Internet as businesses connect to sell their products and services. Academics collaborate over the Internet by exchanging electronic mail. People can converse using Internet phones, send faxes, participate in online chats and bulletin boards, play multiuser games, and experiment with virtual environments.

Special-purpose computers called routers connect the Internet together. As data is forwarded from one place in the Internet to another, it is the routers that make the decisions as to where and how the data is forwarded. The protocols that dynamically inform the routers of the paths that the data should take are called routing protocols. It is the job of these protocols to react quickly to changes in the Internet’s infrastructure, such as transmission lines going in and out of service, routers crashing, changes in network policies, and so on.

Routing is what makes the Internet tick. Although many users of the Internet and the World Wide Web are unaware of the machinery underlying the network applications, routing is an interesting but complicated subject. Routing protocols are sophisticated distributed algorithms that must also be extremely robust to keep a large, decentralized network like the Internet running smoothly.

This book is for students of data communications, TCP/IP network administrators, protocol designers, developers of routing protocol software, and other professionals involved in the design, development, and management of TCP/IP networks. The book is a practical, hands-on description of Internet routing rather than a theoretical treatment. Although we describe how the various protocols were intended to work, we also describe how well the design has translated into practice. Internet protocol design is a practical undertaking itself, with efficiency of implementation often dictating design choices. For this reason, this book gives an in-depth treatment of how a router really works. Instead of just describing the algorithms, the book goes beyond to show how the algorithms are implemented.

We often present ideas in a historical context, showing how Internet protocols have evolved. This is done for two reasons. First, you can learn a lot from the mistakes (and successes) of the past. Second, in order to participate in Internet discussion groups, many of which are dominated by old-timers, it is good to have some context.

This book is not an elementary introduction to TCP/IP and its routing. Instead we assume that you have some familiarity with the TCP/IP protocol suite and some exposure to the basic concepts of routing. These assumptions allow us to explore many of the facets of Internet routing in greater detail than possible in an introductory text.

This book is organized into five parts. Part I sets the groundwork for a discussion of Internet routing. After a brief description of how routing fits together with the rest of the Internet’s protocols, Chapter 1 describes in depth how a router forwards packets. This discussion naturally leads to an explanation of IP addressing and CIDR, as well as of the interaction of hosts and routers. Internet routing protocols are introduced in Chapter 2, beginning with a treatment of the end product of all routing protocols: the router’s routing table. Chapter 2 ends with an overview of the Internet’s routing architecture and the two main routing technologies in use in today’s Internet: Distance Vector and link-state algorithms.

Part II describes the Internet’s OSPF routing protocol. We start in Chapter 3 with an explanation of why the OSPF protocol was developed in the first place. Chapter 4 discusses the basics of link-state routing; Chapter 5, how OSPF behaves over various subnet technologies; Chapter 6, its use of hierarchical routing; and Chapter 7, extensions to OSPF. Each chapter not only describes how OSPF works but also explains why it works that way. We explore the reasons behind OSPF’s design decisions and how the OSPF protocol has evolved to keep pace with the rapidly changing Internet environment. Part II concludes with an OSPF FAQ (Chapter 8).

Part III (Chapters 9 and 10) describes TCP/IP multicast routing, including broadcast and multicast forwarding, the MBONE, and the two distinct types of multicast routing protocols: source-based trees and shared-tree algorithms. As we did with unicast routing, we go further into the subject of multicast routing through the examination of a particular multicast routing protocol: the Multicast Extensions to OSPF (MOSPF).

Part IV covers the configuration and management of Internet routing. The configuration and management of OSPF is explained in detail in Chapter 11. Chapter 12 describes the tools used to monitor and debug routing in a TCP/IP network. For each tool, we describe its use, how it works, and its advantages and drawbacks.

Part V is a comparison of Internet routing protocols. Chapter 13 compares and contrasts the routing protocols in use in the Internet: RIP, OSPF, BGP, IGRP, and IS-IS. In Chapter 14, we examine the available multicast protocols: DVMRP, MOSPF, PIM Dense and Sparse, and CBT.

Following Chapter 14 is an extensive bibliography arranged and numbered in alphabetical order. Within the text, the citation [85], for example, refers to item 85 in the bibliography.

The companion book OSPF Complete Implementation, in keeping with the Internet tradition that reveres “working code” over all else, explores even further the mechanics of Internet routing through examination of a real, working OSPF implementation. The book contains a complete implementation of OSPF on CD. Written in C++, the OSPF implementation is intended to be portable to a wide range of environments. Two sample ports are included: an OSPF routing daemon (called ospfd) for FreeBSD 2.1 and an OSPF routing simulator that can be run on Windows 95. The OSPF implementation has been developed using publicly available tools.

I would like to thank the technical reviewers who improved this book through their thoughtful and timely reviews: Ran Atkinson, Eural Authement, Fred Baker, Howard Berkowitz, Jeffrey Burgan, Joel Halpern, Mukesh Kacker, Robert Minnear, Jim Reid, and W. Richard Stevens. Thanks also to Tim Stoddard and the Arkansas Public School Computer Network (APSCN) for letting me collect OSPF statistics on the APSCN network and use that network as an example of OSPF configuration in Chapter 11, OSPF Management. Thanks to S. Randall McLamb for drawing the figures.

I would also like to acknowledge the help of my editors at Addison Wesley Longman over the long life of this project: Carol Long, Karen Gettman, and Mary Harrington.

And special thanks to my wife, Sonya Keene, who designed the book, edited rough drafts, created the index, and gave encouragement while this book was being written.