SDN: Software Defined Networks

Chapter 8. Network Topology and Topological Information Abstraction

Introduction

Topological information and its availability to applications that wish to utilize what amounts to rare gold in the networking world has long been available to a select few applications. These applications had to satisfy a few important criteria. First, they needed to be fluent in one of many routing protocols. And secondly, they had to be allowed to join in routing in the literal sense: they had to directly attach to a routing network—this took more than just racking up a server and plugging it into a network.

It often required the application to go through the same security and other quality assurance hurdles that any other piece of networking gear in that network had to. This often took literally years of testing to complete. And when that moment arrived when the application was allowed on the network, its grasp of network topology was then still limited to that of active topology—that which was used to actually route or steer traffic at that instant in time.

Inactive or dormant topological information is generally not required for routing calculations and thus is not carried in any routing protocol exchanges. This information is needed by some applications, as explained later in this chapter, and so in order to gain that information an application still had to manually locate it using out-of-band means such as the command-line interface or other management protocols.

One final downside to traditional ...