0
$\begingroup$

I'm looking at collision detection in communication protocols, in particular Carrier sense multiple access with collision detection (CSMA/CD). According to what I've read on Wikipedia, a collision seems to cause the wave of the outgoing signal and the wave of the incoming signal to overlap. Thus, the signal has values with more amplitude than allowed for a "1-bit". This could be used as a working collision detection algorithm.

However, what's the use of the "Jam signal", as explained in the same Wikipedia article, then? It says that the CRC has to be faulted, but we already have a collision detection algorithm.

$\endgroup$
2
$\begingroup$

Wave overlaps need not cause just 1-bits with a too high amplitude. They could also flip bits around, or nullify each other. Thus error correction is not as easy as you describe and one actually has to use some kind of checksumming like CRC to detect errors reliably.

$\endgroup$
2
  • 1
    $\begingroup$ Thanks! Just to be sure I got it: A sender checks if, while it sends its data, new data comes in. If this isn't the case, no collision occured. But if it is, we append a jam signal to the message we're still sending. The receivers will read the appended jam signal and so are noticed of the collision. Is this correct? $\endgroup$ – Johannes Jan 15 '13 at 21:05
  • $\begingroup$ Right. Go for a networking textbook (or look for on-line lecture notes) for further details and analysis. $\endgroup$ – vonbrand Jan 27 '13 at 3:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.