TCP vs UDP Error-Checking

Question

I've bounced upon two articles that describe in-depth on how TCP and UDP connections work. However there is some kind of inconsistency. Where one article tells me that UDP does no Error-Checking at all, the other article mentions that UDP Does some sorth of error-checking but no recovery or corrections. I would like to know which of the two is the most correct?

Thanks in advance!

Solution 1

UDP does have a 16-byte checksum. As Gnouc mentioned, this has some error-detection properties, but doesn't do recovery (by re-sending the packet like TCP will). The article was probably referring to the fact that UDP doesn't have any positive acknowledgement that the packet was received. (i.e. there is no detection if the packet got lost.)

http://en.wikipedia.org/wiki/User_Datagram_Protocol

Solution 2

Yes, the second article is right. The UDP protocol also has error-checking but doesn't have any error-recovery.

• Error-detection: detect error occurs on the frame ( FCS does that )
• Error-recovery: using sequence of bytes to detect error occurs, and if it happen, resend it.