RE: IPS-All: WG Last Call on "iSCSI Naming and Discovery"

[Jim Hafner <hafner@almaden.ibm.com>]

Rob,

I think this is a reasonable and acceptable change.  Thanks.

Jim Hafner

Sent by:        owner-ips@ece.cmu.edu

To:        ips@ece.cmu.edu
cc:        
Subject:        RE: IPS-All:  WG Last Call on "iSCSI Naming and Discovery"



Hello,

One aspect of the Naming and Discovery draft that I find misleading lies in
the sentence in Section B.3

"The gateway may manufacture its own  iSCSI Names, or use those provided by
the real devices."

Now B.3 deals with non-iSCSI devices, so does the phrase "those provided by
the real devices" refer to non-iSCSI devices providing an iSCSI Name? I read
"those provided by the real devices" as the native identifiers (i.e. FC
WWN). This may lead to the conclusion that a Gateway can use the "eui." form
of the iSCSI name in a fairly straight-forward manner. Unfortunately, we
know is not true (at least for the case where multiple gateways can see the
same device on a SAN). So - I suggest a caveat is warranted, like

"It is the responsibility of the gateway to ensure the uniqueness of any
iSCSI name it manufactures. The gateway may need to account for multiple
gateways having access to a single real device".

Now - of course I would REALLY like to see something done at the naming
level that would allow gateway vendors to preserve some part of the identity
of the real device in a non-proprietary fashion. But at least alerting
readers to the problem would be a good step.

Regards,
Rob.