Making Distributed Rate Control using Lyapunov Drifts a Reality in Wireless Networks

In recent years the literature on wireless networks has been enriched by several theoretical results that have developed new mathematical frameworks for design of optimal cross-layer protocols [1]. A particularly appealing stochastic network optimization approach that yields simple distributed algorithms for dynamic scenarios is based on the use of Lyapunov drifts [2]. The Lyapunov drift based techniques described in [2] provide for distributed rate control based purely on local observations of neighborhood queues. They guarantee the stability of the system and also provide mechanisms to achieve optimization with respect to given utility functions. While attractive on theoretical grounds, to our knowledge these techniques have yet to be implemented in real wireless networks 1. One reason it has not been easy to translate the Lyapunov drift methodology from theory to practice is that it has been primarily developed under the assumption of a time-slotted system, implying a TDMA MAC. TDMA-based wireless networks are generally harder to implement due to the challenges associated with time-synchronization, particularly across multiple-hops.