QoS per TCP connection

Jozsef Kadlecsik kadlec@blackhole.kfki.hu
Tue, 2 Oct 2001 10:15:12 +0200 (CEST)


On Sun, 30 Sep 2001, Sampsa Ranta wrote:

> You wrote that piece of code to track TCP window.. Do you by any
> change have a piece of code that allows adjustments to that window size.

No, nothing.

> I just learned (in university, something good comes out of those, too =)
> that window size is the thing that tells us how much data there can be in
> transit and adjusting it could be used as to limit bandwidth used by a
> single TCP connection atleast on link with known bandwidth.
> In other words, this could be used as very simple shaper for a single TCP
> connection, which would be good and simple way.. In case we also had
> byte counter in conntrack entry, it would be possible to start limiting
> data rate if the TCP connection was used to transfer data enough.
> What do you think?

First of all, the window size could only be shrinked. However, shrinking
the window is a bad practice - not illegal, but discouraged by the RFC.

But it could be done: the actual window sizes are already stored, one
would store the modified window sizes as well. However, it would
then belong to mangle and not conntrack because of the packet-modification
and because it'd need a table for the rules which connection and at what
conditions should be adjusted.

E-mail  : kadlec@blackhole.kfki.hu, kadlec@sunserv.kfki.hu
WWW-Home: http://www.kfki.hu/~kadlec
Address : KFKI Research Institute for Particle and Nuclear Physics
          H-1525 Budapest 114, POB. 49, Hungary