[storm] WG Last Call: iSCSI drafts

[<[e-mail]>]

This message announces a storm Working Group Last Call on the following two iSCSI drafts:

			iSCSI Protocol (Consolidated)
	  	    draft-ietf-storm-iscsi-cons-03.txt

       Internet Small Computer Systems Interface (iSCSI) SAM
                 draft-ietf-storm-iscsi-sam-03.txt

Please send all technical comments to the storm mailing list ([e-mail]).
Editorial comments may be sent directly to the draft editors:
- iSCSI Protocol (Consolidated): Mallikarjun Chadalapaka <[e-mail]>
- iSCSI SAM (new features): Frederick Knight <[e-mail]>

This is going to be a long Working Group Last Call because we need feedback from implementers.
Nonetheless, it would be good to have some WG Last Call review input before the storm WG meeting
in two weeks in Quebec City (Tuesday, July 26, 1710-1810).

This Working Group Last Call will end at midnight, US Eastern Time on Sunday, August 21.  Not
counting the IETF meeting week, that's about a 4 week last call time period.  

The Protocol draft is enormous, and mostly unchanged, so I don't expect implementers to read
the whole thing.  Please direct implementers to the "Summary of Changes" section on pp.23-24
of the Protocol draft and ask them to look at the additions in the SAM draft.

I will be contacting the EMC and VMware iSCSI implementers, plus the Linux iSCSI maintainers.
I believe that Mallikarjun and Fred have Microsoft and NetApp covered.  Feedback from other
implementers is strongly encouraged - anyone who can arrange for feedback from another iSCSI
implementation, should please do so, and send a note identifying the implementation to the
list or directly to the WG chairs:

	 Tom Talpey <[e-mail]> and David Black <[e-mail]>.

I'll get this started with several comments on the SAM draft (these comments are as an
individual, not as a WG co-chair):

(1) The draft title should change.  The SAM acronym needs to be expanded, and the
	title needs to convey that there are new iSCSI features being added.

(2) It makes sense to use the IANA registry for the new iSCSIProtocolLevel key as the
	source of the revision numbers for the SCSI version descriptors for iSCSI - see section
	D.8 in Annex D of SPC-3 or the current draft of SPC-4, and see the specification of
	the INQUIRY command for their usage.  In order to do this, the draft needs a couple
	of changes:
	- The value range for iSCSIProtocolLevel needs to be limited to 0-31.  We'll
		be assigning the values 0, 1 and 2 to start with.
	- For assignment of new values, review by a T10 expert reviewer should be added to
		the requirement for a published standards track RFC.  How to state this process
		requirement will be a topic for discussion at the storm WG meeting in Quebec City.

Plus one on the Consolidated Draft - it's enormous ... while I don't think that the main
draft can be split (getting this all into one document was a goal), we should consider whether
the Annexes can be put into a separate document.  Unfortunately, Annex D on SendTargets is
definitely normative, and Annex E on clearing effects may be normative, plus all the annexes
were originally included in the main draft because they are useful to implementers.  This is
for discussion on the list and in Quebec City.

Thanks,
--David
----------------------------------------------------
David L. Black, Distinguished Engineer
EMC Corporation, 176 South St., Hopkinton, MA  01748
[phone]             FAX: [phone]
[e-mail]        Mobile: [phone]
----------------------------------------------------