[Iccrg] Heresy following "TCP: Train-wreck"

[Lachlan Andrew]

On 05/04/2008, Dirceu Cavendish <dirceu_cavendish at yahoo.com> wrote:
>
> About BitTorrent type of applications, it would be nice to have some sort of fairness concept between these types of application, and across to other applications. But then, assuming there are N types of applications in the Internet (which I dont have to say N can be large), one would need to provide fairness definitions for N(N-1) cases.

We don't need N(N-1) cases.  Just as we use money as an abstraction to
avoid exchange rates for the k(k-1) goods which could be bartered, we
could have a "congestion currency".  That is Kelly's approach.  His
work was based on instantaneous rates etc, so would need to be
extended, but shows a way forward.

We might need a few, representing different QoS parameters;  Kevin
Tang has been working on extending that framework to networks with
multiple "congestion prices" at each link.

Another issue is how to deal with fairness which isn't just "dividing
a resource".  For example, consider one flow wanting high throughput,
which is achieved with queueing, and another sharing that bottleneck
wanting low jitter, which is achieved by keeping buffers empty.  What
is the "fair" behaviour?

Cheers,
Lachlan

-- 
Lachlan Andrew  Dept of Computer Science, Caltech
1200 E California Blvd, Mail Code 256-80, Pasadena CA 91125, USA
Ph: +1 (626) 395-8820    Fax: +1 (626) 568-3603
http://netlab.caltech.edu/lachlan