[e2e] Why do we need TCP flow control (rwnd)?

[Stefanos Harhalakis]

On Friday 11 July 2008, Christian Huitema wrote:
> I suspect that any process that results from human behavior is very likely
> to exhibit much more variability than predicted by a Poisson model. That's
> certainly true for phone traffic as well.

  While I'm not related to statistics I believe that it is almost impossible 
to (statistically) model a group A that is affected by groups B and C without 
first modeling B and C. As you said, Internet traffic is affected by many 
external sources like human behaviour, random facts etc, which are by 
themselves extremely complex. Since we've not modeled (yet?) those external 
sources (regarding their Internet behaviour influence) we most probably won't 
be able to accurately model Internet traffic. The only exception that I can 
think of is that we could be extremely lucky and the whole thing would just 
follow a common model.