[Iccrg] Re: TCP requirements

[Lachlan Andrew]

Greetings Keshav,

On 12/10/06, S. Keshav <keshav at uwaterloo.ca> wrote:
> However, over the long term, you may achieve
> max-min-fair-share, but *not* Pareto optimality.

> I think the confusion arises from the fact that there are two time-
> scales of interest.

The key is max-min-fair-*share*.  That is max-min on a reduced
feasible region, like the "constrained Max-min" concept introduced in
the paper "Understanding XCP".  To me that is not "max-min", but
you're right that it is often said that non-Pareto protocols like XCP
are max-min fair.

Anyway, I didn't mean to detract from Bob's and your original point
that seeking fairness at all costs is not good, and should be
discussed.

Thanks for your patience,
Lachlan


-- 
Lachlan Andrew  Dept of Computer Science, Caltech
1200 E California Blvd, Mail Code 256-80, Pasadena CA 91125, USA
Phone: +1 (626) 395-8820    Fax: +1 (626) 568-3603