RE: Towards Consensus on TCP Connections

[julian_satran@il.ibm.com]

IEEE 1394 is a legitimate contender for the local area - and provided that
prices are kept low
by the competition in the consumer space - may make iSCSI and interesting
control scheme
for a wide variety of devices  - including but not limited to storage.

Julo

Peter Johansson <PJohansson@acm.org> on 12/08/2000 19:28:37

Please respond to Peter Johansson <PJohansson@acm.org>

To:   IP Storage <IPS@ece.cmu.edu>
cc:    (bcc: Julian Satran/Haifa/IBM)
Subject:  RE: Towards Consensus on TCP Connections




At 11:14 AM 8/11/00, Douglas Otis wrote:

>Firewire does not scale and has a limited reach.

Doug, I take exception to the first part of your statement, both because I
believe that IEEE 1394 does scale and for another reason (see below). As
for limited reach, I agree that point-to-point connections measured in the
hundreds of meters are not adequate for a MAN / WAN. But I also don't think
that consideration is relevant to iSCSI (from the protocol perspective),
whose purpose is to permit a diversity of media solutions appropriate to
each user's needs.

>Firewire also places scatter/gather on the drive together with direct
access.

I think you're confusing SBP-2 (which does locate the scatter / gather
chore in the disk drive) with IEEE 1394 (which is agnostic about such
matters). SBP-2 cannot be "native" iSCSI any more than can FCP!

It's my understanding that the charter of iSCSI is to be transport neutral.
That is, iSCSI is supposed to run over anything---copper, fibre, tautly
stretched string---that supports Internet protocol. Is this belief in
error?

If that assumption is correct, then when it comes time to compare and
contrast the cost and efficiency of different media, IEEE 1394 ought to be
in the matrix along with Ethernet, FC, UTP-5, etc., etc.


Regards,

Peter Johansson

Congruent Software, Inc.
98 Colorado Avenue
Berkeley, CA  94707

(510) 527-3926
(510) 527-3856 FAX

PJohansson@ACM.org