Talk:Message switching

The first sentence ("message switching was the precursor of packet switching") seems to indicate that message switching and packet switching are two distinct paradigms. The second sentence ("Message switching systems are nowadays mostly implemented over packet-switched or circuit-switched data networks") seems to suggest that packet switching is an implementation of message switching. Can we clarify the relationship between message switching and packet switching? (My guess is that the second sentence may be in error, but I'm not a telecommunications expert, so I don't know.) —Caesura(t) 17:02, 11 February 2006 (UTC)

The layout with the diagram is not really good, but I can't fix it - or not yet - at any rate.

If someone else wants to have a go - get the diagram to the left, or the centre, that's fine by me! My simple attempts didn't work. David Martland 10:40, 1 October 2007 (UTC)

Caesura

Message switching is suitable for short messages, but as should be clear from the section added, is not really practical for very long messages (e.g files, images etc), and not feasible at all for streaming data. Both have routing across networks, and experience delay due to propagation and transmission delays. With packet switching the routing delay is limited by the size of the packets used (typically 1.5 kbytes for packets which originate on Ethernets, or up to 64kbytes for IP packets), and in any case with modern systems, which may use cut through routing, the delays can be minimised further. Additionally, high data transfer rates used nowadays reduce the transmission delays, but when message switching was commonly used, transmission delays could have been quite long. Both packet switching and message switching have end to end delays which depend on the number of nodes along the path - typically 30 or less. It is also possible that with modern networks that queueing delays at intermediate nodes are more significant than transmission delays. David Martland 10:48, 1 October 2007 (UTC)