[e2e] Is a non-TCP solution dead?

[Cannara]

The problem, Jon, remains that the grail of "TCP friendliness" is insufficient
because it isn't all the traffic by any means.  Being friendly to TCP, because
TCP is trying to be 'friendly' to some IP nets, too often makes TCP unfriendly
to real users & applications.  This is characteristic of kludges -- pushing
problems farther away from their sources and avoiding real solutions.

Alex

Jon Crowcroft wrote:
> 
> In missive <DAC3FCB50E31C54987CD10797DA511BA027E11E9 at WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.c
> om>, "Christian Huitema" typed:
> 
>  >>I suggest people read:
> 
>  >>S. Shenker, "Making Greed Work in Networks: A Game-Theoretic Analysis of
>  >>Switch Service Disciplines," in SIGCOMM Symposium on Communications
>  >>Architectures and Protocols, (London, UK), pp. 47-57, Sept. 1994.
> 
> Yes, and
> Rate control in communication networks: shadow prices, proportional fairness and stability
> F. P. Kelly, A.K. Maulloo and D.K.H. Tan (Statistical Laboratory, University of Cambridge)
> Journal of the Operational Research Society 49 (1998), 237-252.
> http://www.statslab.cam.ac.uk/~frank/rate.html
> 
> together, these make for a pretty solid argument for "tcp friendliness"
> which i have yet to see any effective refutation.
> 
> j.