Cooperative Fair Bandwidth Scaling in Contention-based Wireless Networks using Time Token Bucket

Applications with Quality of Service (QoS) requirements such as networked control systems are not well supported by today's off-the-shelf wireless standards such as IEEE 802.11 (WiFi). Specifying a range of minimally required base QoS and preferred QoS combined with dynamic scaling in this range can help to support QoS while improving network utilization. However, scaling is typically happening in a greedy and therefore unfair way, with each node trying to maximize its throughput. In this paper, we present a cooperative algorithm for dynamical cooperative bandwidth scaling using a time token bucket approach. We show that our approach improves bandwidth utilization in a fair way. Furthermore, it dynamically reacts to topology changes, e.g. due to node mobility, as well as to traffic of uncooperative (external) nodes. Finally, it achieves stable bandwidth usage of participating nodes, even in networks with external traffic. We evaluate our approach in an IEEE 802.11 testbed.