Layered congestion control for scalable video coding based on the efficient bandwidth inference

Scalable video coding allows greater granularity adaptation at bit level to the time-vary heterogeneous networks. In this paper, we propose a congestion control algorithm based on the bandwidth estimation techniques. In order to cooperate with H.264/MPEG-4 AVC SVC extension which can achieve fine granularity of scalability, we design an adaptation strategy to smartly switch to an efficient bandwidth inference mode during the channel probing period according to the one way delay and jitter profile so that the subscription decision of SVC layers can be made quickly to better utilize the network resource. In case of the unavoidable network congestion, we unsubscribe scalable video layers according to the recently received throughput instead of only dropping one layer at a time to rapidly accommodate the streaming service to the channels. The simulations show that the proposed algorithm converges fast to the available bandwidth and can adapt to the fluctuated network efficiently.