Re: iSCSI: DLB's [T.6] 2.3 iSCSI Session Types

["John Hufferd" <hufferd@us.ibm.com>]

I also believe that those extended discovery functions belong in SLP and
iSNS, not in extending the Discovery Session.  The Discovery Session was
intended for the simple lower end iSCSI network, that did not have SLP or
iSNS.  And in that end of the market there is no reason to have long lived
Discovery sessions.  Also this will cause vendors to invent ways to work
with the base discovery session, that were unintended and thus cause
additional problems, and not focus on getting interoperable support for SLP
and iSNS.

Each of our approaches, "discovery sessions", SLP, and iSNS, generally
cover different network markets, and this segmentation is rather easy to
explain.  I would prefer to continue this focus, or we will confuse the
market and ourselves and never be interoperable.

.
.
.
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com


Prasenjit Sarkar/Almaden/IBM@IBMUS@ece.cmu.edu on 07/11/2002 08:15:38 AM

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


To:    Mark Bakke <mbakke@cisco.com>
cc:    Black_David@emc.com, ips@ece.cmu.edu, Julian Satran/Haifa/IBM@IBMIL,
       owner-ips@ece.cmu.edu
Subject:    Re: iSCSI: DLB's [T.6] 2.3  iSCSI Session Types




I'm not sure what Mark means by "we", but I am opposed to any dynamic
discovery mechanisms for SendTargets because of my belief that these
mechanisms does not belong here. Therefore I vote against Noop and Async
messages,

Sincerely,

   Prasenjit Sarkar
   Research Staff Member
   IBM Almaden Research
   San Jose




                      Mark Bakke
                      <mbakke@cisco.com        To:
                      Black_David@emc.com
                      >                        cc:       Julian
                      Satran/Haifa/IBM@IBMIL, ips@ece.cmu.edu
                      Sent by:                 Subject:  Re: iSCSI: DLB's
                      [T.6] 2.3  iSCSI Session Types
                      owner-ips@ece.cmu
                      .edu


                      07/11/2002 08:21
                      AM





Actually, we want to be able to send Async Message and Noop In
from the target, and allow the initiator to send Noop Out, since
these are more in the iSCSI session maintenance category of
messages.

There's definitely no reason to allow Scsi Command / Response,
Task Management Command / Response, Data In/Out, Snack, and
R2T on a discovery session.

--
Mark

Black_David@emc.com wrote:
>
> Julian
>
> > I can agree with MUST but I will resist adding any notification.
> >
> > It is just the wrong thing to do.
>
> I don't care - getting to a MUST with a list of allowed PDUs that
> forbids all others from the current MAY that allows everything is
> what matters to me.
>
> At the moment, I've only seen Ayman ask for Async Message to be
> allowed on discovery sessions - does anyone else want to see that
> PDU allowed on discovery sessions?
>
> Thanks,
> --David
> ---------------------------------------------------
> David L. Black, Senior Technologist
> EMC Corporation, 42 South St., Hopkinton, MA  01748
> +1 (508) 249-6449            FAX: +1 (508) 497-8018
> black_david@emc.com       Mobile: +1 (978) 394-7754
> ---------------------------------------------------

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