[tcpm] Adding Support for Stronger Error Detection Codes in TCP for Jumbo Frames

["Biswas, Anumita" <[e-mail]>]

Hello,
 
An internet-draft on "Support for Stronger Error Detection Codes in TCP
for Jumbo Frames" has been submitted. 
 
It can be found at
http://www.ietf.org/id/draft-anumita-tcpm-stronger-checksum-00.txt
<http://www.ietf.org/id/draft-anumita-tcpm-stronger-checksum-00.txt> 
 
Any feedback, discussion and comments would be highly appreciated.
 
Many thanks,
Anumita Biswas
Abstract:
   There is a class of data serving protocols and applications that
   cannot tolerate undetected data corruption on the wire.  Data
   corruption could occur at the source in software, in the network
   interface card, out on the link, on intermediate routers or at the
   destination network interface card or node.  The Ethernet CRC and the
   16-bit checksum in the TCP/UDP headers are used to detect data
   errors.  Most applications rely on these checksums to detect data
   corruptions and do not use any checksums or CRC checks at their
   level.  Research has shown that the TCP/UDP checksums are catching a
   significant number of errors, however, the research suggests that one
   packet in 10 billion will have an error that goes undetected for
   Ethernet MTU frames (MTU of 1500).  Under certain situations, "bad"
   hosts can introduce undetected errors at a much higher frequency and
   order.  With the use of Jumbo frames on the rise, and therefore more
   data bits on the wire that could be corrupted, the current 16-bit
   TCP/UDP checksum, or the Ethernet 32-bit CRC are simply not
   sufficient for detecting errors.  This document specifies a proposal
   to use stronger checksum algorithms for TCP Jumbo Frames for IPv4 and
   IPv6 networks.  The Castagnoli CRC 32C algorithm used in iSCSI and
   SCTP is proposed as the error detection code of choice.