End-to-End Tomography and Congestion Control on Multicast Based Videoconferencing

Congestions in many-to-many videoconferencing are usually caused by multiple senders, the congestion information feedback, bottleneck link judge and sending rate adjust are much different from one-to-many multicasting. In this paper, we propose one multicast congestion control scheme based on end-to-end tomography (ET-MCC) for many-to-many videoconferencing. The end-to-end packet loss rates are measured and aggregated for judging bottleneck links through tomography approach. One performance server is adopted to aggregate the packet loss rate and do tomography with low overhead. The rates of senders that generate traffic on bottleneck links are equally adjusted to their acceptable value to realize congestion control with sender-fairness. By simulations and applying the protocol to live videoconferencing of CERNET-MVC, ET-MCC is proved to be valid and effective.