RE: iSCSI: URL scheme

["Douglas Otis" <dotis@sanlight.net>]

John,

Sorry, but it would seem you are blind to the problem.  I was trying very
hard to make a point.  You can not use symbolic names within a transport.
You should exclude these symbols and configuration from the transport.  If
you wish to define names, make it a class object within an LDAP server.  As
you again fail to address the basic concerns, I still wonder.

Doug



> -----Original Message-----
> From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu]On Behalf Of
> John Hufferd/San Jose/IBM
> Sent: Wednesday, October 04, 2000 11:02 AM
> To: ips@ece.cmu.edu
> Subject: RE: iSCSI: URL scheme
>
>
>
> Douglas Otis,
> Perhaps you miss addressed your note, I do not believe that I
> either stated
> what you claim or even believe what you say I believe.  Please
> recheck your
> source.
>
> What I wanted was for folks to start from the basics and tell us step by
> step what they think they need in the way of a name, why,  how they would
> get it, and how they would use it (including how it would react to current
> network equipment and current  Storage Controllers --which may have an
> iSCSI interface.  I have strong doubts that some of the ideas I have been
> hearing are needed or reasonable.
> .
> .
> .
> John L. Hufferd
>
>
>
> "Douglas Otis" <dotis@sanlight.net> on 10/04/2000 09:00:31 AM
>
> To:   John Hufferd/San Jose/IBM@IBMUS, <ips@ece.cmu.edu>
> cc:
> Subject:  RE: iSCSI: URL scheme
>
>
>
> John,
>
> The problem that I see with your present scheme that is causing the
> confusion and difficulty is from trying to create an amalgam of
> IP and SCSI
> addresses.  Although you may view URLs as extensible and perhaps
> capable of
> handling both IP and SCSI, the load created will be horrendous.  In judo,
> they teach you how use inertia in your favor.  In this case, you are
> attempting to change the course of the entire world using a new symbolic
> router.  I am not as clever as you.  I only wish to translate firmly
> establish binary fields that address devices.  Yes, I can translate these
> and perhaps even bridge some of them.  I can not parse every
> packet looking
> for a symbolic address and then route this packet dynamically in both SCSI
> and IP domains with your new breed of symbolic router, name and
> authentication server. On top of that, you wish to see everything
> extracted
> from this single connection for authentication and configuration.  We end
> up
> doing a great deal of looking at these packets and then a great deal of
> deciding what to do with them.  None of this is productive work.  You
> develop a class 2 control for a class 3 world for additional trouble.  I
> would hope that you want a simple solution and not one sure to thwart all
> efforts.
>
> Your truly,
> Doug
>
>
>
>
>
>
>