Talk:Remote collision

The origin of the name "remote"
In my opinion the article describes the remote collision correctly but it essentially describes an effect, not the cause of the remote collision.

In fact, the term "remote collision" comes from times when repeaters (and hubs) were used on networks. If a collision occured on one segment connected to the repeater, the collision fragment together with the jam signal was transmitted by the repeater to the other segment. The collision was local at the segment where it occured in the first place, and remote at the other segment where it was propagated by the repeater.

The local and the remote collision are similar in that the collision fragment appears as a frame that is shorter than the minimum size and whose FCS is incorrect. The most notable difference between local and remote collision is that local collision violates the coding scheme while the remote collision does not. The signal on the cable may experience illegal levels (and perhaps timing) due to superposition of the signals coming from colliding NICs. However, these illegal signal levels will not get transmitted through the repeater - as the repeater regenerates the signal itself, it can never generate something that violates the coding scheme. Therefore, the remote collision indeed appears as a frame fragment that is shorter than minimum frame size, has an incorrect FCS but that does not violate the coding scheme.

This is what I have known about it - if there is anybody knowing better please let us all know. Paluchpeter 23:24, 13 January 2007 (UTC)