Re: TCP/IP CRC question

[Fernando Gont <fernando () gont com ar>]

At 01:15 06/10/2004 -0500, Jorge Mendez Bonini wrote:

> The Ethernet Frame contains a field known as FCS which contains a CRC.
> Now, my question is:
> If the CRC is generated from DATA field among others, What's the point of 
> using checksums on the upper protocolos (IP checksums etc..)?

First, the link-layer checksum "protects" data only on agiven network segment.
Also, some of the links in the path may not implement error detection. 
Therefore you need to implement this function yourself in upper layers.
The IP checksum protects the IP header on a end-node to end-node basis. 
However, it does not protect the payload. The TCP checksum protects data on 
an end-to-end basis (strictly speaking, not really).

Finally, CRCs and checksums catch different type of errors. Checksums catch 
single bit errors (which usually happen at routers' memories), while CRCs 
protect data against burst error (continuous bits in the data stream are 
corrupted).


--
Fernando Gont
e-mail: fernando () gont com ar || fgont () acm org