Non-Saturation and Saturation Analysis of IEEE 802.11e EDCA with Starvation Prediction

Paal E. Engelstad and Olav N. Osterbo

This paper was very clarifying for me in many ways. It uses Markov Chains and presents an extension to take into account AIFS.

TCP Muro

There are two problems with current TCP:
- The slow start in high bandwith-delay links.
- The persistent p2p flows that consume most of the Internet's bandwith.

I believe that if the growing of the congestion window could be modified to increase also as a function of time at the beginning, the two problems would be solved. Or, at least, alleviated.

Game-theoretical approach to CSMA/ECA

It might be possible to show that a deterministic backoff after successes and a random backoff after failures is the nash equilibrium of a game.

Or maybe the nash equilibrium is a bit more complex: using a random backoff only if the collision resulted from a random backoff. This leads to a quick convergence protocol, a bit unfair by those that not find their place (slot) immediately.