A counterexample in congestion control of wireless networks

This paper studies the interaction between TCP congestion control and wireless interference. One of the triumphs of wireline network research of the last decade has been the casting of the Internet congestion control problem within an optimization framework based on utility functions. Such an approach has provided a sound theoretical understanding of the underlying stability and fairness issues, as well as a post-facto justification of the scalability and stability of TCP-like additive-increase multiplicative-decrease (AIMD) algorithms. This paper provides counterexamples showing that the same result cannot be extended to wireless networks, at least not in a straightforward manner. The fundamental difference is that wireless networks are of a broadcast nature. There is no strict notion of a 'link', since transmissions from nearby nodes interfere with each other. We consider a fairly general model of interference in wireless networks, and present a counterexample of a wireless network in which the congestion control mechanism has an unstable equilibrium point at the desired fair solution. ns-2 simulations of this counterexample manifest an oscillatory throughput behaviour that is orders of magnitude worse than the corresponding wired networks. Surprisingly, this oscillatory throughput behaviour appears to be fairly typical of simulations in wireless networks, with almost all randomly chosen network simulation examples manifesting it. This loss of stability leads us to suggest that perhaps TCP should be modified for use in wireless networks, and that a cross-layer redesign of wireless TCP and MAC is needed to explicitly account for the effects of the wireless nature of interference.