Re: [tcpm] TCP tuning

[Michael Welzl <michawe at ifi.uio.no>]

Hi,

> > http://trac.tools.ietf.org/group/irtf/trac/wiki/ICCRG
> >
> > IMO, I would encourage Google to take their work there and get  
> > feedback
> > there first. TCPM is a place where mature, vetted ideas should be
> > considered, preferably (IMO).
>
> Sure we can do that and will love to hear feedback from ICCRG. But our
> preference is IETF simply because we consider a simple, modest
> increase of the initial window to be more of an engineering project  
> than
> a research problem. The former is more about engineering tradeoff  
> between
> cost/risk, benefit and TIMELINESS than a research project to find  
> the best,
> elegant, risk free solution with little or no time-to-market concern.

Without trying to get into a debate on engineering vs. research, I would
say that the research community's perception of TCP research is that,
indeed, such a change is a valid research question. That's not because
of a need to find the "best, elegant, risk free solution", but because
of the need to make sure that it's a solution that's safe to deploy.

This is indeed at odds with needs for timeliness - but your need for
that can't be the basis for lowering the bar about significant changes
to TCP in the IETF.


> We are aware of all the great research work being conducted at ICCRG
> on speeding up slow-start. But none of the proposals seems ready for
> IETF anytime soon. I'd be happy to be proven wrong here! (Well to be
> more correct a modest increase of initial window may have already been
> studied at ICCRG. In that case I'd rephrase it to be the closest to
> engineering consideration IMHO.)

The proposals that have been made in the ICCRG context were not
there because this is the most, or only, appropriate way to do things
in ICCRG - they were there because they were proposed. I would
love to see proposals for more "IETF-ready" approach there.
I don't think that a modest increase of the initial window has been
considered in the ICCRG context, and it would fit very well.

Cheers,
Michael