Cut-through switching
From Free net encyclopedia
In computer networking, cut-through switching is a switching method for packet switching systems, wherein the switch starts forwarding that frame (or packet) before the whole frame has been received, normally as soon as the destination address is processed. This technique reduces latency through the switch, but decreases reliability.
In packet switched networks such as Ethernet, pure cut-through switching can only be used where the speed of the outgoing interface is less than or equal to the incoming interface speed.
Cut-through routing in IP networks presents some problems since the IP checksum in the packet header is supposed to be checked by every router in the path. Since the checksum of a packet cannot be checked until the entire packet has been received, the cut-through router is at risk of forwarding a packet with an incorrect checksum. Provided that there are other routers in the path which are not doing cut-through routing, or that the end system is correctly verifying checksums, this should only result in the occasional loss of a small amount of traffic capacity.
(The above statement concerning checksum calculation is wrong. IP checksum checks only the header bytes, not the whole packet. The reason for not adopting cut-through routing is among others queueing backlog)
Cut through routing was one of the important features of IP networks using ATM networks since the edge routers of the ATM network were able to use cell switching through the core of the network with low latency at all points. With higher speed links, this has become less of a problem since packet latency has become much smaller.
Cut-through switching is very popular in InfiniBand networks, since these are often deployed in environments where latency is a prime concern, such as supercomputer clusters.