A Novel Multiple Time Scale Congestion Control Scheme for Multimedia Traffic

摘要

A novel two time scale congestion control scheme is proposed in this paper, which is called Adaptive Wavelet and Probability-based scheme (AWP). On normal time scales of 20-200 ms, it is a smooth and TCP-friendly congestion control; on large time scales of 1-5 second, it detects persistent shifts in overall network contention and modulates the rate control exhibited on normal scale. AWP exploits self-similar property of network traffic and predicts future network available bandwidth to enhance performance, and it is adaptive to network environment changes through an auto-correction algorithm. With comparable experiments among TFRC, TCP and AWP over NS platform, AWP has been proved to be able to greatly improve the quality of service(QoS) of multimedia data transmission while avoiding congestion collapse on the network.