This chapter forms the heart of this book and represents my original impulse to write about UNIX routing.
It is almost impossible to tame and reliably operate rapidly changing environments and topologies solely with the use of static routes. This especially is an issue in Internet service provider (ISP) networks with large numbers of dynamically created routes via PPP-connections from dial or Digital Subscriber Line (DSL) customers. However, the main purpose of dynamic routing is providing high availability in case of node or link failures. Dynamic routing protocols were designed to adapt to topology changes and distribute routing information within an autonomous system, based on an underlying algorithm.
This chapter discusses the two families of Interior Gateway Routing Protocols (IGRPs) and provides extensive lab scenarios using Routing Information Protocol (RIP), Open Shortest Path First (OSPF), and a quick introduction to Intermediate System-to-Intermediate System (IS-IS), which most likely is rather exotic or even alien to UNIX and even some network folks. Based on the insight of the previous chapter, this chapter extends the view to how dynamic routing protocols populate the kernel routing table and takes a closer look at making proper choices for deployment based on essential characteristics of these protocols as well as service and topology requirements. This chapter also introduces the concept of areas?virtual segmentations of network realms?a specialty of link-state routing protocols. Special topics such as route exports and redistribution, Equal-Cost Multi-Path (ECMP) issues, and traffic-engineering extensions conclude the chapter.