Chapter 8. Scaling

Scalable may well be one of the most overused words in the network designer's lexicon. It applies to protocols, to software, to operating systems, to hardware architecture, and to networks. Scalable means, simply enough, the ability of the relevant entity (select one from the list just given) to get much bigger than it presently is without reducing performance, stability, or accuracy; making your customers angry; or getting you fired.

You have already encountered in this book a number of features that make OSPF and IS-IS scalable from the smallest to the largest networks. The most prominent feature for both protocols is areas. By dividing a network into multiple areas, you can control in each area the scope of flooding, the size of the link state database, and the complexity of the SPF calculations. By bounding these three fundamental link state functions, you constrain the demand OSPF and IS-IS put on router memory, router processing cycles, and link bandwidth, respectively.

You have also encountered features or extensions that can increase scalability, such as configurable refresh timers and increased metric sizes. This chapter delves further into features and extensions that can increase the scalability of the protocols.