[Iccrg] "Slow Start" design team

[John Leslie]

Michael Welzl <michael.welzl at uibk.ac.at> wrote:
> 
> A design team should have a very narrow focus, its own
> mailing list, and be led by someone who is willing and
> able to push the whole thing along.

   Agreed.

> The first concrete proposal on the table is (somewhat
> unsurprisingly  :)  ) a design team on Slow Start,
> led by Dirceu himself (which I, not he, suggested).
> I think it's a nice start and definitely worth trying.
> 
> To join the list, send an email to:
> iccrg-slowstart-ctl at ndrc.kyutech.ac.jp
> with "subscribe YOUR NAME" as the body of the message.

   I'm not sure a Design Team should be quite that open,
though I'd certainly welcome frequent interaction with the
main iccrg list.

   In any case, I'm holding off on joining that list...

> The first question that should be discussed in this
> list (not here - let's try and move this discussion there
> already) is what exactly the team should focus on:
> 
> * enhancements to the end of slow start (like Dirceu
>   presented at our meeting)
> 
> * different slow start enhancements (like the paper from
>   Injong's group at PFLDNet)
> 
> * something else?

   I strongly recommend nailing down the problem that
Slow-Start is supposed to solve. Without firm agreement
on that, I'm afraid we'll get a random-walk problem.

> * is this only about TCP, or also for other protocols?

   That might be the wrong question. We should start with
a description of what we're trying to fix; then work on
whether fixing it within TCP is even the right place to
start.

--
John Leslie <john at jlc.net>