RE: iSCSI:Target Centric ISID assignments

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

Sandeep,

The vendor-unique id is yet another scheme to add another layer of
identifying information (beyond InitiatorName and ISID).  As I've mentioned
in other posts, I don't see structural reason for it (unless the ISID space
is just too small, then we can argue about where to put and what to put in
the additional space we create).

However, one problem with the vendor-unique suggestion is the same problem
we in N&D wrestled with, namely, not every component has an identifiable
vendor and that vendor have an unique identifier to go with it.  It's true
for HW components, but all those fancy SW versions of iSCSI initiators
don't.  That's what lead us to the iqn naming scheme for InitiatorName.


Jim Hafner


sandeepj@research.bell-labs.com (Sandeep Joshi)@ece.cmu.edu on 09/09/2001
09:47:44 am

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


To:   Julian Satran/Haifa/IBM@IBMIL, ips@ece.cmu.edu
cc:
Subject:  RE: iSCSI:Target Centric ISID assignments



Agree with the extra responsibility added on the target.

But one point here ..the target may not be monolithic
but one assumes it would atleast be "monogenic" (single-vendor)
thereby enabling it to disallow multiple nexuses being
started with the same <initiatorName,ISID>

The monogenic property may not hold for initiators so
a scheme which works without HBA cooperation is preferred
over one which requires cooperation.

The schemes presented by Jim in a previous email (a windows
registry counter) still assume all HBAs follow that
assignment model.

The ISID namespace seems too small to partition among
vendors.  If it were possible to add a vendor-unique
identifier to the login command/phase (and maybe nexus),
we may solve the problem without burdening the target.

-Sandeep

> Blowing away or not is again with us. A careless guy will set always the
> blow-away bit.
> You have just moved the management responsibility to the target under the
> assumption that the target
> is "monolithic". But, as you well know, this is not true for many
targets.
> You have to have an allocation scheme at
> target and a garbage collection.
>
> I suggest we give this to the ND team and let them debate for a while
> before getting back to us.
> We may want to have them look at all aspects of SSID allocation and use.
>
> Julo