RE: Performance of iSCSI, FCIP and iFCP

["Somesh Gupta" <somesh_gupta@silverbacksystems.com>]

I am more hopeful on the point of adapting the algorithm
as environments change. The TCP gurus are conservative but
not totally inflexible.

-----Original Message-----
From: Victor Firoiu [mailto:vfiroiu@nortelnetworks.com]
Sent: Thursday, January 18, 2001 1:34 PM
To: somesh_gupta
Cc: IPS Reflector
Subject: Re: Performance of iSCSI, FCIP and iFCP


Somesh,

I was looking more at the comparison between the FCIP case of multiple
FC connections (all connections between two FCIP gateways) per TCP
connection and the iFCP case of one/few TCP connections per FC
connection (thus multiple TCP connections per pair of iFCP gateways).

I think that, in general, one TCP connection per storage connection is
closer to "one/two TCP connections per user" model, and thus 
easier to defend as being "fair", compared to multiple TCP connections
per storage connection.  I agree, this is quite a fuzzy quantification,
but is as precise as RFC 2914 and 2309 get in this matter.

Regarding modifing TCP congestion control as you suggest, I am again
very skeptical.

Victor




Somesh Gupta wrote:
> 
> Victor,
> 
> I just assumed that you were familiar with the iSCSI specification
> of multiple connections per session - which is essentially as you
> describe in the second last paragraph - there is one big "flow"
> between the initiator and the target (called an iSCSI session)
> which is then striped across multiple TCP connections - and I
> think people assumed that your statements were in support of
> such a proposal.
> 
> Having said that, the math still holds. And also the fact that
> fundamental limits to the throughput of a single TCP connection
> really do require a change to TCP congestion avoidance (esp when
> mistaking link error for congestion) formulae.
> 
> If the link error rate is such that errors occur faster than time
> to recover (or of that order), TCP will have to accomodate that
> somehow. I don't think the TCP gurus are inflexible in this area
> - PAWS, window scaling, and SACK are all example of TCP accomodating
> the characteristics of the underlying network.
> 
> Somesh
>