RE: "Lower overhead" iSCSI

[Black_David@emc.com]

Some comments from an IETF perspective:

(1) The IETF does not standardize protocols that are intended only
	for use in private networks, especially when the topology of
	those private networks has to be limited to ensure proper
	operation.
(2) The best intentions to restrict use of a protocol to
	such private networks often go awry - if the protocol is
	seriously useful, it will get used in other places.
(3) IETF will only consider standardization of an iSCSI over UDP
	protocol when TCP-equivalent/friendly congestion control
	is included.  This is not negotiable - don't even ask.
(4) The NFS experience should be considered - NFS was originally
	specified to run over UDP, but most NFS servers in use
	today run over TCP due to its better behavior on real
	networks.  The TCP vs. UDP performance deltas for NFS
	are not that large even when the CPU is saturated.

Thanks,
--David

> -----Original Message-----
> From: Dan_McConnell@Dell.com [mailto:Dan_McConnell@Dell.com]
> Sent: Tuesday, January 29, 2002 5:30 PM
> To: ips@ece.cmu.edu
> Subject: "Lower overhead" iSCSI 
> 
> 
> Let me start off by saying that I am interested in doing "iSCSI" protocol
> over UDP.  Now I realize that this is an old issue and will probably start
> some "religious" battles, but let me state the scenarios before I receive
> death threats.  The planned environment that this will go into is a small
> one with say 10 servers connected through a "non-blocking" switch to the
> storage device (ie no routers, gateways, etc...just direct point-to-point
> connections).  This is assuming that the switch is really non-blocking and
> hopefully implements flow control or pause frames.  So technically all you
> should have to worry about is port/device contention.  However, when you
> think about it...this is similar to FC.  FCP runs on class 3 FC which is a
> non-reliable transport protocol such as UDP and handles contention, also
> some of the early "SAN interconnect" guys are doing this today with
> relatively good performance and few issues.
> 	The attempt here is to maintain low CPU utilization at high
> performance rates.  While I realize that these TOE devices are moving
along
> rapidly, there are some situations where they are not feasible, such as a
> blade server environment (no PCI slots, and no real estate/power available
> for onboard TOE).  Worst case scenario is that a packet is dropped or
> received out of order and the ULP (SCSI) must resend the cmd/data sequence
-
> still no data lost, just a temporary performance hit.
> 	So my question is:  is this feasible?, and why not implement an
> "iSCSI" protocol layer that can run over TCP or UDP(though I realize it
> won't be considered "standards compliant")?
> 
> Thanks,
> Dan
>