Restrained medium access control on adversarial shared channels

Abstract

We study the fundamental problem of utilization of the channel shared by stations competing to transmit packets on the channel. In their turn, packets are continuously injected into stations' queues by an adversary at a rate at most ρ packets per round. The aim of the distributed medium access control algorithms is to successfully transmit packets and maintain system stability (bounded queues). We further restrain algorithms by introducing an upper bound k on the allowed number of active stations in any given round. We construct adaptive and full sensing protocols with optimal throughput 1 and almost optimal throughput $1-ϵ$ (for any positive ϵ), respectively, in a constant-restrained channel. On the opposite side, we show that restricted protocols based on schedules known in advance suffer from a substantial drop of throughput, at most $\min \{\frac{k}{n},\frac{1}{\log n}\}$. We compare our algorithms experimentally with well-known backoff protocols.