BGP is what keeps the Internet flowing. ISPs use BGP to inform each other which IP addresses go where. Are you making the most of BGP at your ISP?

by Patricia Fusco Managing Editor ISP-Planet [November 22, 2002]

The Internet is a "network of networks," which is one of reasons why it has scaled so well to its current size. The Internet has evolved into what it is today because its creators built it upon a foundation of rigid routing protocols. The Border Gateway Protocol (BGP) is one of the ways that routers keep up with ever-changing routing tables. The protocol plays a key role in assuring the smooth transit of data over the Internet. ISPs use BGP to express a routing policy of their own, so that many different users can be routed to many different destinations, each under the direction of a single autonomous system.

Iljitsch van Beijnum has long been intrigued by computer networking. Even before dropping out of high school, he was intrigued by the protocols that computers use to communicate.

After a brief stint in college, van Beijnum left to work at a small ISP. He started to learn about Cisco routers and the countless quirks of interesting protocols running them—most notably BGP. He has configured the protocol on single-router networks, networks with several hundred Cisco routers ranging from the slowest to the fastest available—even in multi-vendor environments with BGP running on Cisco and Juniper routers, Extreme switches, and FreeBSD hosts running GNU Zebra. When he's not working as a network engineer for UUNET Netherlands, van Beijnum freelances as a networking consultant. Which is how he came to "write the book" on BGP.

The book, titled BGP-Building Reliable Networks with BGP, is for anyone interested in running BGP to create reliable connectivity to the Internet. It caters especially to smaller ISPs with related interests in interconnecting (peering) with other networks and providing BGP transit services. Specific prior knowledge isn't required for reading this book, but some exposure to basic networking theory will be helpful.

The book contains just about everything you need to know to run BGP for regular IPv4 routing in all but the largest network scenarios. In chapter one, van Beijnum discusses how ISP networks connect together to form the Internet. There is a short overview of TCP/IP design principles, as well as an explanation as to why there must be inter-domain routing protocols as well as intra-domain routing protocols.

The second chapter takes a long, hard look at IP addressing and the inner workings of BGP, including the multi-protocol extensions and the BGP route selection algorithm. In chapter three, van Beijnum discusses the physical side of the network—higher availability through redundancy, router hardware, and network topology. Chapter four reviews the various types of IP address space—along with their limitations—and introduces the Routing Registry system.

In chapter five, van Beijnum changes directions and takes a look at the practical elements of getting started with BGP. He explains in great detail how to configure external BGP (eBGP) to a single ISP and how to determine whether your address block shows up on routers in other networks. From there, van Beijnum goes on to discuss how to use a second router to connect to another ISP, and how to configure internal BGP sessions. Chapter six discusses traffic engineering and how to take advantage of having two connections to the Internet by optimizing traffic flow. Chapters five and six include Routing Policy Specification Language (RPSL) examples for several different routing policies.

Chapter seven takes a look at the best way to secure access to your routers, the use of Telnet versus SSH, and inherent weaknesses in the software. But the main thrust of this chapter is about protecting BGP against problems created by other networks—either intentionally or unintentionally. This includes extensive information on using BGP to deflect distributed denial of service (DDoS) attacks.

Chapter eight introduces the day-to-day requirements that inter-domain routing imposes on a Network Operations Center (NOC) and how to manage routine BGP operations. Chapter nine takes a look at troubleshooting physical and datalink problems and goes into great detail discussing inter-domain routing snafus and reachability problems.

The last three chapters examine everyday issues that impact ISPs and some of the design challenges inherent to BGP. Included is a discussion of BGP peer groups, use of loopback addresses, and scaling—as well as some of the pitfalls of providing transit services. In the final chapter, van Beijnum makes a business case for exchanging traffic with other networks, explains how to connect to an exchange point, and discusses routing issues associated with connecting to several exchange points.

Since the book is aimed at readers with varying levels of BGP knowledge, it allows both experts and beginners to be at ease with the presentation. The appendixes provide detailed information about Cisco configuration basics, binary logic, netmasks and prefixes, and notes on IPv4 address space. There is also a detailed glossary of BGP terminology, just in case you get lost in the alphabet soup of routing jargon.

As mentioned, specific prior knowledge of BGP isn't required for reading this book, but some exposure to basic networking theory is recommended. If you need to broaden your understanding of routing principles, or are trying to get a handle on load balancing or the benefits of peering agreements for your ISP, van Beijnum's book is a great place to start. It will make you think about routing not only as a way to build an efficient network, but also as a means to fulfilling your business goals, too...important goals like providing your ISP's customers with reliable, redundant data transport services.

—End