RE: TargetAlias/InitiatorAlias text commands

["KRUEGER,MARJORIE (HP-Roseville,ex1)" <marjorie_krueger@hp.com>]

You are describing something that is exposed to the user via a management
interface to your implementation.  There is no value of these constructs *to
the iSCSI protocol operation*.  I don't deny the value to a user, but a
management interface provides the iSCSI contact point to a user, and can
provide the association of iSCSI names to an "alias string" within an
implementation environment.  There is nothing similar contained in any
transport protocol specification that I'm aware of, and I don't think that's
"a mistake", but a design of specifying only what's necessary to protocol
operation.  

Alias names are appropriate to the MIB, or some other management definition,
not the iSCSI protocol.

Marj

> -----Original Message-----
> From: Mark Bakke [mailto:mbakke@cisco.com]
> Sent: Tuesday, July 24, 2001 2:21 PM
> To: Jim Hafner
> Cc: ips@ece.cmu.edu
> Subject: Re: TargetAlias/InitiatorAlias text commands
> 
> 
> 
> I don't agree.  Although TargetAlias and InitiatorAlias are not
> used within the SendTargets response, they ARE specified as being
> sent during the login phase.  Basically, the initiator would send
> its InitiatorAlias whenever it sends InitiatorName, and the target
> would send back TargetAlias within the login response.  This enables
> initiators and targets to have non-unique, user-friendly names
> apart from their iSCSI names, and a way to communicate them.  These
> can then be used within a user interface as specified in the
> NDT document's alias section.
> 
> I really do think that these add value; gateways such as ours
> provide access to multiple targets, which the user is allowed to
> name.  These names are not necessarily globally-unique, but they
> are meaningful to the user.  A method to communicate these names
> provides a method to give a user that warm, fuzzy feeling that
> he or she has found the right drive.
> 
> Keep in mind that some target and initiator names may be based
> on EUI-64, and will basically be a bunch of hex digits.  This
> was not very friendly or usable with Fibre Channel; why should
> we make the same mistake?
> 
> --
> Mark
> 
> Jim Hafner wrote:
> > 
> > Marj,
> > 
> > I concur.  I don't see any value in these keys either, at 
> least within the
> > protocol.
> > 
> > Jim Hafner
> > 
> > "KRUEGER,MARJORIE (HP-Roseville,ex1)" 
> <marjorie_krueger@hp.com>@ece.cmu.edu
> > on 07-24-2001 01:38:35 PM
> > 
> > Sent by:  owner-ips@ece.cmu.edu
> > 
> > To:   "Ips Reflector (E-mail)" <ips@ece.cmu.edu>
> > cc:
> > Subject:  TargetAlias/InitiatorAlias text commands
> > 
> > As outlined in David Black's consensus statement on the 
> SendTargets command
> > 
(http://www.pdl.cmu.edu/mailinglists/ips/mail/msg05119.html), the target
> records returned in response to a SendTargets request will not include
> TargetAlias, in response to the expressed desire to limit the SendTargets
> functionality to the basic description of the target iSCSI device.
> 
> There are references to "aliases" sprinkled throughout the iSCSI protocol
> document, but this construct doesn't add any value to the iSCSI protocol
> model or functionality.
> 
> Since the TargetAlias and InitiatorAlias text commands do not contribute
> any
> functionality to the iSCSI protocol, are not used in any iSCSI-related
> logic
> within an implementation, and are basicly a "label" more appropriately
> administered by a management tool, I propose they be removed from the
iSCSI
> protocol spec.
> 
> Comments?
> 
> Marjorie Krueger
> Networked Storage Architecture
> Networked Storage Solutions Org.
> Hewlett-Packard
> tel: +1 916 785 2656
> fax: +1 916 785 0391
> email: marjorie_krueger@hp.com

-- 
Mark A. Bakke
Cisco Systems
mbakke@cisco.com
763.398.1054