SES: Stable and Efficient Solution for Rate Control and Spectrum Allocation in Wireless LANs

The technology of DSA (Dynamic Spectrum Access) will be adopted in the next generation of wireless network to maximize spectrum utilization. The fundamental challenge of applying DSA into wireless network is how to provide efficient spectrum allocation and rate control for DSA users. In this paper, we propose a Stable and Efficient Solution, called SES, to optimize the rate control and spectrum allocation in a typical enterprise or campus wireless LAN. SES is decoupled into three independent algorithms for initial spectrum allocation, rate control, and dynamic spectrum allocation. A distributed algorithm adopting the technique of Lyapunov optimization is designed to tune the rate control of client and AP (Access Point). Furthermore, to maximize spectrum utilization, we introduce a novel notion, called "Virtual Queue of Sub-network", to evaluate the realistic spectrum demand of individual AP radio, based on which a centralized algorithm has been developed to dynamically optimize the spectrum allocation. Theoretical analysis shows that SES can make optimal decision on rate control and spectrum allocation, and stabilize network with explicit performance bounds. Simulation results demonstrate that SES can achieve up to 90% improvement in throughput and 50% reduction in queue length, as compared to the solution of fixed channel.