Towards Consensus on TCP Connections

[Black_David@emc.com]

Let's see if we can make some progress towards
consensus here.  I see two open issues in the
discussion on connection issues:

(A) Should iSCSI require a TCP connection per LUN?
(B) Should iSCSI have a session abstraction that
	binds multiple TCP connections into one
	iSCSI connection?

On the first issue, (A), I have only seen one
individual (David Robinson) advocating the "yes"
position.  Therefore, I believe there is rough
consensus for "no" as the answer to (A).  This
does not forbid building a system in which each
iSCSI target has a single LUN, but also allows
multiple LUNs per target.  In RFC 2119 terms,
I think that this will come down to:
- Initiators MUST support multiple LUNs/target
- Targets MAY support multiple LUNs/target
One of the things that needs to be done to the iSCSI
draft is to clarify the requirements by reference
to RFC 2119 and use of the capitalized terms found
therein.

NB: Now I'm going to give the game away by explaining
what I just did in IETF terms.  As WG co-chair, I have
expressed my opinion that rough consensus exists; that
opinion may be incorrect since WG co-chairs are not
infallible by any means.  If anyone other than the one
individual noted above believes that "yes" is the best
technical answer to (A), they should speak up on the list,
especially if I've missed a message advocating "yes"
from someone else in reviewing the traffic on the list.

Despite Steve Byan's recent change of opinion, I do not
believe that we have reached rough consensus on (B).  Part
of the reason for this is that I'm not sure everyone has
clearly understood the consequences of a "no" on (B).
Let me start from Costa's simple explanation:

> There are two ways in which multiple connections are used in iSCSI:
> 
> 1) Multiple simultaneous TCP connections for bandwidth
> 	
> 2) Multiple TCP connections for fault tolerance/recovery
> 	(i.e. when one TCP connection in a session dies, 
>          another one starts up)

It is an established practice in the storage industry
to achieve both 1) and 2) via the use of multiple
SCSI connections, and evidenced by numerous products
that do this for host access to storage.  Hence a
"no" answer to (B) does not preclude the use of
multiple TCP connections for bandwidth or fault
tolerance/recovery as long as one is willing to use
multiple SCSI connections to do so.

So, to resolve issue (B), we need to discuss whether
there are compelling reasons for doing 1) and/or 2)
within a single SCSI connection, as opposed to across
multiple SCSI connections.  This specific issue needs
further discussion -- I invite everyone to have at it ...

--David

---------------------------------------------------
David L. Black, Senior Technologist
EMC Corporation, 42 South St., Hopkinton, MA  01748
+1 (508) 435-1000 x75140, FAX: +1 (508) 497-6909
black_david@emc.com  Cellular: +1 (978) 394-7754
---------------------------------------------------