[Iccrg] Incentives for ECN

[Lachlan Andrew]

Greetings John,

On 15 July 2011 22:27, John Leslie <john at jlc.net> wrote:
>
>   I would suggest that the essential feature of ECN is that it delivers
> a congestion signal _sooner_ -- so a CC algorithm which _initially_
> backs off more quickly upon receipt of an ECN mark, but after a few RTTs
> perhaps is more aggresive than vanilla-TCP upon recognizing a loss is
> a concept worth research.

It would have been nice if ECN had originally been specified to signal
congestion sooner, but since it wasn't, it isn't clear how to migrate
to a system that does, without breaking things.

As Bob Briscoe has argued
<http://www.ietf.org/mail-archive/web/ledbat/current/msg00479.html>,
it causes problems for algorithms (such as standard TCP) that don't
distinguish between ECN and loss.  If ECN occurs sooner, it causes
ECN-aware flows to be starved, which is the wrong incentive (as well
as being too harsh).

Cheers,
Lachlan

-- 
Lachlan Andrew  Centre for Advanced Internet Architectures (CAIA)
Swinburne University of Technology, Melbourne, Australia
<http://caia.swin.edu.au/cv/landrew>
Ph +61 3 9214 4837