Re: FCIP/iFCP : Guarantee In-Order delivery for FC N/NL_ports

[Matt Wakeley <matt_wakeley@agilent.com>]

YP,

You are correct that most people in *this* group understand the problems. 
However, *this* group cannot invent new TCP options.  That is controlled by
the end2end group, and *they* do not understand the problems (or if they do,
they are unwilling to help solve them).

-Matt


Y P Cheng wrote:
> 
> > If I read you correctly, you are asserting that there is a fundemental
> > problem in the design on the TCP *protocol* which prevents it from taking
> > advantage of a 10G/100ms network. If so what exactly do you see as a
> problem?
> 
> I have been very careful in stating the problem is in implementation.  The
> protocol is fine -- as the more and more I learn about the TCP options.  I
> think more new options such as framing, marking, NCK, SACK, and RDMA can be
> defined and placed in login negotiation to allow the TCP protocol to run
> well in 10G/100ms network.
> 
> > Since we cannot (ips WG) cannot change TCP how should an IPS protocol
> > work around this problem while still being friendly with other protocols?
> 
> I do believe many people in this group do understand the problems.  It new
> options that facilitate the 10G/100ms network are negotiated, we can still
> be friendly with older implementation which does not support the new
> options.  In such case, we just run slow.
> 
> > > If everyone agrees that this group can put iSCSI, iFCP, and FCIP
> together by
> > > assuming the current TCP implementations having all the solutions,
> please
> > > let me know.
> >
> > Conversely, if you feel that this group is designing to the TCP
> > implementations instead of the protocol, please let us know.
> >
> >       -David
> 
> I did sense that some people in this group were worry about compatibility
> with older implementations and reluctant to discuss or add new TCP options.
> In general, I do believe lots of people are quite up to speed, pun
> unintended.  My statements made in herein previous postings were saying that
> two iSCSI or FCIP adapters of same kind -- with same TCP implementations --
> should be able to run the new options.
> 
> Y.P.