RE: TCP/IP CRC question

["Ted A" <arcturous () hotmail com>]

Also, remember that the stack was built with the intention of being 
completely modular. By allowing each layer to be responsible for it's own 
duties, you don't have a common point of failure; IE: Bad checksum as a 
result of routing will also mean bad data even if the data is good.
By having the seperate points of responsibility, troubleshooting is more 
reliable.




> -----Original Message-----
> From: Jorge Mendez Bonini [mailto:jlmb () cableonda net]
> Sent: Tuesday, October 05, 2004 11:15 PM
> To: security-basics () securityfocus com
> Subject: TCP/IP CRC question
>
>
> I've been reading about the TCP/IP protocol stack
> (TCP/Ilustrated Vol1
> by Richard Stevens) lately and since Ethernet is quite common
> nowadays
> almost all link layer examples refer to it (I also checked
> Douglas Comer
>   Internetworking with TCP/IP vol1).
> I tried searching RFC but found it very time consuming without good
> results (maybe I'm not used to it...yet)
>
> 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..)?
>
> Thanks for your time.
>