[Tmrg] Proposal to increase TCP initial CWND

Michael.Scharf at alcatel-lucent.com (SCHARF, Michael) Sun, 18 July 2010 10:16 UTC

From: Michael.Scharf at alcatel-lucent.com (SCHARF, Michael)
Date: Sun, 18 Jul 2010 12:16:37 +0200
Subject: [Tmrg] Proposal to increase TCP initial CWND
In-Reply-To: <AANLkTil937lyUzRvUtdqd2qdl9RN7AZ-Mo_cT-dtmqXz@mail.gmail.com>
References: <AANLkTil937lyUzRvUtdqd2qdl9RN7AZ-Mo_cT-dtmqXz@mail.gmail.com>
Message-ID: <133D9897FB9C5E4E9DF2779DC91E947C038F0954@SLFSNX.rcs.alcatel-research.de>

> The basic TMRG-related issue was:  If a data centre changes a 
> TCP parameter for *its* flows, how can it estimate the 
> degradation caused to *other* flows sharing resources with 
> these flows?  Google is awash with data, but it all shares 
> the bias of coming from their data centres.  What models are 
> appropriate to infer the impact on the unmeasured flows?

I am not sure whether there is a simple way how to estimate the impact
on other flows without somehow measuring them (e. g., on the
client-side).

In the paper http://www.springerlink.com/content/6078217660g68170/, as
well as in follow-up work that is still under submission, I played with
simulations in which a certain share of flows use a fast startup
schemes, and I studied the impact on the other flows (e. g., a potential
slow-down). The simulation scenario implemented a subset of
draft-irtf-tmrg-tests-02.

However, the results of such studies depend a lot on the characteristics
of a potential common bottleneck and its load. (If someone is interested
in this work, please contact me privately.)

Michael