Re: [tcpm] TCP tuning

[Jerry Chu <hkchu at google.com>]

On Wed, Feb 3, 2010 at 7:17 AM, SCHARF, Michael
<Michael.Scharf at alcatel-lucent.com> wrote:
>> This topic seems to be gaining momentum, and the WG should
>> take some time considering if there is work here for it.
>
> IMHO there could indeed be room for increasing the initial window. If
> most data transfers continue to be smaller than 3 MSS, a larger initial
> window would not necessarily cause harm, as it is seldomly used. Still,
> this would speed up those data transfers that currently suffer from
> Slow-Start.

Actually our data points to the contrary - the average web object and page
size have been rising steadily. E.g., the majority of our search responses
no longer fit in 3 MSS these days.

>
> In the past, I have played a little bit with an initial window e. g. of
> 10 MSS, and that worked surprisingly well.

Good to know!

>Also, one could combine a
> larger initial window with other mechanisms that would reduce the
> aggressiveness of the flow startup. Maybe it is indeed time to think
> about the Slow-Start once again.

We are planning to submit an I-D on raising initcwnd to 10MSS before
3/1 to be included in the upcoming IETF meeting.

Jerry

>
> Michael
> _______________________________________________
> tcpm mailing list
> tcpm at ietf.org
> https://www.ietf.org/mailman/listinfo/tcpm
>