RE: The iSCSI Plugfest - update for Wednesday

["Julian Satran" <Julian_Satran@il.ibm.com>]

No - only sending good status on Data-In PDU.  And again this did not
change from draft 00.  Julo

"Barry Reinhold" <bbrtrebia@mediaone.net> on 19-07-2001 15:23:47

Please respond to "Barry Reinhold" <bbrtrebia@mediaone.net>

To:   Julian Satran/Haifa/IBM@IBMIL, ips@ece.cmu.edu
cc:
Subject:  RE: The iSCSI Plugfest - update for Wednesday




Julian,
     Am I to understand from this that iSCSI intends to support both (1)
sending
of status information in iSCSI data PDUs and also sending of data in iSCSI
response PDUS?

>-----Original Message-----
>From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu]On Behalf Of
>Julian Satran
>Sent: Thursday, July 19, 2001 6:27 AM
>To: ips@ece.cmu.edu
>Subject: Re: The iSCSI Plugfest - update for Wednesday
>
>
>
>Status and Data where (and) are part of the standard.
>They are important to reduce the protocol overhead and performance of the
>software implementation
>is substantially reduced without it.
>
>It is clearly stated in several places (1.2, the Data-In PDU) etc. and has
>never changed (from version 0).
>
>As for the plugfest - congratulations - great work.
>
>Julo
>
>
>
>
>Please respond to "Barry Reinhold" <bbrtrebia@mediaone.net>
>
>To:   "ISCSI" <ips@ece.cmu.edu>
>cc:
>Subject:  The iSCSI Plugfest - update for Wednesday
>
>
>
>
>The plugfest process has settled down to the detailed process of isolating
>implementation bugs. A significant percentage of the implementors are now
>able to connect and perform I/O operations - though this is often achieved
>by first compiling with -D "company Name" switches to handle the different
>issues that have already been discovered.
>
>New Issues: A detailed posting of the issues to date will (or has) be done
>by Bob Russell. In brief  only one additional standards issues came to
>light
>that impacted interoperability in the days testing:
>
>1. The current wording of the standard appears to allow for the
>transmission
>of IO data in response frames. This is actually being done in a few
>implementations. Although this was perhaps a vision at the start of the
>iSCSI effort there is little support for this. Most implementors were not
>prepared to handle this.
>
>Progress:
>
>Test progressed to round 17 and it appears that it is highly likely that
>developers will get a chance to test against all other parties by the end
>of
>the week.
>
>Two vendors were able to establish a multi connection session, getting
>through the login process and the start of IO.
>
>Two other vendors established an iSCSI session with header digests,
>completing the security phase of the login process.
>
>Some generic problem areas that come up:
>
>0 Length data frames
>Target initiated text negotiation
>Response information in data PDUs.
>SCSI compatibility issues
>
>There was a UNH iSCSI consortium meeting at the end of the day
>
>1. Overview of UNH iSCSI consortium
>2. Discussion on dates/time for next plug fest
>3. Election of steering team
>
>People flowing out at 7:00 -- sign that things are pretty good.
>
>
>
>Barry Reinhold
>Principal Architect
>Trebia Networks
>barry.reinhold@trebia.com
>603-868-5144/603-659-0885/978-929-0830 x138
>
>
>
>