Re: iSCSI: 2.2.6. Naming & mapping

[meth@il.ibm.com]

Peter,
By "transport-dependent" I assume you mean dependent on the SCSI transport
(in our case iSCSI, similar to the WWN used in FCP) rather than on TCP. You
seem to be suggesting that we should define in iSCSI a way to name and
discover LUs  without performing SCSI Inquiry commands. (Do I understand
you correctly?) Can you please give a scenario of the type of case you have
in mind with the benefits of introducing this new feature? In your
scenario, would the ability to name the particular LU be crucial, or would
it still be sufficient to simply name the target device and then perform
Inquiry on the target device?

Could you also suggest how you think the naming of a particular LU should
be specified?

Thanks,
- Kalman



Peter Johansson <PJohansson@ACM.org> on 18/09/2000 01:03:10

Please respond to Peter Johansson <PJohansson@ACM.org>

To:   IP Storage <IPS@ece.cmu.edu>
cc:    (bcc: Kalman Meth/Haifa/IBM)
Subject:  Re: iSCSI: 2.2.6. Naming & mapping





<... deleted ...>

Jim Hafner's observation (that mandatory presence of a unique ID, on a per
LU basis, in EVPD appears to be destined for adoption within SPC-2) only
reinforces that part of the discussion with which I already agree. It is
desirable to have command set-dependent methods to determine the unique ID
for a LU. Matter closed?

It might also be desirable to have transport-dependent mechanisms to
specify unique IDs for those LUs whose existence is discoverable by
transport-dependent methods. I think this is still open to discussion.

<... deleted ...>




Regards,

Peter Johansson

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

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

PJohansson@ACM.org