RE: IPS and SCTP

[Black_David@emc.com]

Brian,

> Sure.  Perhaps a start would be (for me) to write an
> individual submission extension draft for running, say,
> iSCSI over SCTP.  Would that be of interest to this WG?
> What would be the best timing for such a submission?

I can't speak for the WG as a whole, but that sounds like
a reasonable course of action to me.  Getting a first
version of your draft submitted far enough in advance
of the Atlanta meetings so that people will have time to read
and consider it (i.e., well before the "just prior to the
deadline" draft deluge) ought to allow the WG the opportunity
to decide whether to make it an official work item in Atlanta.

A few iSCSI over SCTP things to start with:
- Header digests can be eliminated since SCTP now uses CRC32C,
	and should result in a visible simplification of error
	recovery.
- Unless SCTP has changed to allow load balancing as part of
	its multi-homing/failover support, iSCSI sessions will
	need to support multiple SCTP sessions per connection
	for load balancing.
- SCTP's ability to avoid head-of-line blocking should be useful
	when immediate delivery of iSCSI commands is requested.

Enjoy,
--David
---------------------------------------------------
David L. Black, Senior Technologist
EMC Corporation, 42 South St., Hopkinton, MA  01748
+1 (508) 249-6449            FAX: +1 (508) 497-8018
black_david@emc.com       Mobile: +1 (978) 394-7754
---------------------------------------------------