RE: SCSI URL scheme [WAS: Re: iSCSI: 2.2.6. Naming & mapping]

["John Hufferd/San Jose/IBM" <hufferd@us.ibm.com>]

Peter Johansson,
Not exactly.

The iSCSI layer does NOT deliver a CDB to a LU.   And believe it or not
this is the important point; it delivers the CDB to the SCSI layer.  It is
the SCSI layer that defines how the LU is addressed, not the transport, and
it is the SCSI layer that delivers the CDB to the LU.

If you have not had the chance to review Jim Hafner's notes in this area,
you might want to do so, they are, in general, very elucidating.


.
.
.
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
(408) 256-0403, Tie: 276-0403
Internet address: hufferd@us.ibm.com


Peter Johansson <PJohansson@ACM.org>@ece.cmu.edu on 10/06/2000 05:51:13 PM

Sent by:  owner-ips@ece.cmu.edu


To:   IP Storage <IPS@ece.cmu.edu>
cc:
Subject:  RE: SCSI URL scheme [WAS: Re: iSCSI: 2.2.6. Naming & mapping]



At 01:00 PM 10/3/00, John Hufferd/San Jose/IBM wrote:

>.. I think you [Howard Green] are talking about major changes to the CDB
>structure, and that should be deferred to the T10 committee.

I find myself inadequate to keep pace with the effluence of the IP storage
reflector, so please forgive me if I've missed some subtle point in
Howard's proposal and John's reply (above).

If I understood, Howard was suggesting that the usage of LUNs be bypassed
in favor of direct use of unique identifiers possessed by logical units.
This DOES NOT affect CDB structure. The presence of a LUN in some CDBs is
vestigial and may be ignored.

How to specify a LUN (and what it's maximum size may be) is a matter for
the transport "container" that gets the CDB to the LU. This is entirely
within the domain of iSCSI.




Regards,

Peter Johansson

Congruent Software, Inc.
98 Colorado Avenue
Berkeley, CA  94707

(510) 527-3926
(510) 527-3856 FAX

PJohansson@ACM.org