through the use of a silicon-switching cache located on the Silicon Switch Processor (SSP) on
some Cisco 7000-series routers.
algorithm and the optimized structure of the optimum-switching cache. Optimum switching is
available only on routers equipped with a Route/Switch Processor (RSP).
or other interface cards that can receive route information from the master RSP to make their
own autonomous, multilayer switching decisions. Distributed switching supports very fast
throughput because the switching process occurs on the interface card.
these flows at the same time that it applies services. NetFlow switching also lets a network
manager collect data on network usage to enable capacity planning and to bill users based on
network and application resource utilization. The data can be collected without slowing down
the switching process.
periphery of a network to enable features such as traffic accounting, QoS functionality, and
security, and to use an even faster switching mode in the network's core. At the network's core,
the switching mode should forward packets based on easily accessible information in the packet
and generally should not spend time applying services.
quickly across large backbone networks and the Internet. Rather than relying on the caching
techniques used by classic switching methods, CEF depends on a forwarding information base
(FIB). The FIB allows CEF to be much less processor-intensive than other Layer 3 switching
methods because the FIB tables contain forwarding information for all routes in the routing
tables (whereas a cache contains only a subset of routing information).
are characterized by sessions of short duration to multiple destination addresses. CEF became