Internet Protocol Small Computer System Interface (iSCSI) Cyclic Redundancy Check (CRC)/Checksum Considerations

Abstract

In this memo, we attempt to give some estimates for the probability of undetected errors to facilitate the selection of an error detection code for the Internet Protocol Small Computer System Interface (iSCSI).

Internet Small Computer System Interface (iSCSI) Protocol (Consolidated)

“…Generally, the following factors must be considered before deciding on the proper level of recovery: c) Probability of transport-layer "checksum escape" (message error undetected by TCP checksum --see [RFC3385] for related discussion). This in turn decides the iSCSI digest failure frequency and thus the criticality of iSCSI-level error recovery.…”

The Effectiveness of Checksums for Embedded Control Networks

“…Although Fletcher and Adler checksum error detection properties are almost as good as a relatively weak CRC, they are far worse than good CRCs for some important situations. Fletcher published error detection information in his original paper [28], whereas [21] and [31] present further analysis.…”

“…This is 1-bit less than the Fletcher checksum, which was unexpected since they use an almost identical mathematical basis. (Reference [31] states that the Adler checksum has a higher probability of undetected burst errors than the Fletcher checksum but does not explicitly state that the burst error detection coverage is 1 bit shorter in length.) The reason for this is that Adler checksums use a prime modulo that is less than 2 k À 1, whereas Fletcher checksums use a modulo equal to 2 k À 1, with k being the checksum size.…”

Internet Small Computer Systems Interface (iSCSI)

“…For a discussion about selection criteria for the CRC, see [RFC3385]. For a detailed analysis of the iSCSI polynomial, see [Castagnoli93].…”