Fast Switching and Process Switching

Whether you use fast switching or process switching can affect the way your routes behave. In one project with which I was involved, the administrators wanted to balance traffic across two T1 lines, using EIGRP to perform load balancing. They observed that a daily file transfer was always using a single line, which wasn’t what they wanted. The problem was that EIGRP’s load balancing is session-based, not packet-based. Therefore, once the file transfer started, it was able to use only one of the lines. The solution was to configure the router for process switching rather than fast switching, because process switching is able to load balance on a per-packet basis.

Before looking at this example further, let’s look more closely at the difference between fast switching and process switching and the effect they have on routing.

Fast Switching

When the first packet of a session is going out an interface, a route is selected and placed in a route cache. This route cache entry is used for all packets belonging to this specific destination, which means that all packets belonging to the session take the same route. An entry remains in the route cache until the route cache is flushed, the route changes, or the cache overflows. (When a cache overflows, the entry that was least recently used is removed.)

A session is a communication on a port to a specific host. For example, if Host A is FTPing to Host B, each file is transferred in a single session. Successive ...

Get CISCO IOS in a Nutshell now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.