Re: [Tsvwg] Re: HighSpeed TCP for Large Congestion Windows

>This forum is not the only place where good protocol design is applied 
>or considered. The question is whether it is relevant - at this point - 
>to consider preliminary work.
...
>Keeping the WG informed isn't the same as making this a WG agenda item.
...
>This WG has in the past been fairly conservative about taking on agenda 
>items, notably experimental ones, considering the conservative approach 
>about modifying TCP.
...
>There is 
>insufficient motivation in this case to warrant an exception.

Well, it made most sense to me to ask for the HighSpeed TCP draft
to be considered as an agenda item of tsvwg, and to be considered
for an Experimental RFC, but if tsvwg decides otherwise, then that
is OK by me.

I just thought I would explain my reasons for asking it to be considered
as a WG agenda item:

* HighSpeed TCP proposes a rather significant change in the underlying
TCP response function.  Any such change would have to be globally
agreed-upon.  It seemed advisable to me to bring such a proposal
to the IETF for feedback even in its preliminary stages.

* While several people have conducted evaluations of HighSpeed TCP
using simulations, my own belief is that the "real" evaluations
will have to happen in experiments and in actual deployment.  As
part of this experimentation, HighSpeed TCP has been implemented
in the Linux 2.4.16 Web100 kernel.  It seemed to me that it was
incumbent upon me, at this stage, to bring this to the IETF and to
seek Experimental status.

* I no longer view the world as divided between the "public" and
the "private" Internet, with the IETF only concerned about the
"public" Internet.  I understand that there are ISPs, and parts of
the net operated and controlled by governmental agencies (such as
DOE and NASA in the U.S.) or by other institutions (e.g., schools),
parts that are behind firewalls, etc.  Thus, the fact that
some/most/all(?) high-bandwidth transfers currently are on DOE/NASA/CERN
and similar networks did not seem to me a reason not to bring
something to the IETF.

* A fourth reason to bring this to tsvwg is to get feedback, both
positive and negative, from people who I don't happen to know, and
who I would not otherwise talk with.  Fortunately, this one can
easily be satisfied simply by discussions on the mailing list, or
on some other mailing list, and has nothing to do with being a WG
item or with Experimental status one way or another.

My two bits.

- Sally
_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg