Distributed Alpha-Fair Throughput Aggregation in Multi-RAT Wireless Networks

Today’s wireless devices can be simultaneously connected to multiple communication networks based on different radio access technologies (RATs) such as WiFi, 3G, and LTE. Simultaneous aggregation of each client’s traffic across multiple such RATs or the corresponding base stations (BSs) can substantially improve the quality of experience for each client and improve the overall network utilization and efficiency.Our goal in this paper is to design distributed resource allocation algorithms that can be independently executed by each BS and achieve alpha-fairness among the clients. In particular, we derive a simple water filling based solution and study its theoretical aspects such as convergence, optimality, evolution, and convergence time. We also characterize its performance through a comprehensive multi-RAT simulator, and show that it achieves superior performance to alternative distributed solutions due to its fast and low-overhead operation.