Multiple Description and Layered Coding for Video Streaming Applications

[[abstract]]The objective of this project is to develop a novel layered multiple description video streaming system well suited for peer-to-peer (P2P) networks. In a P2P system, each peer node can be both the information provider and receiver. Therefore, as compared with the basic client-server model for video streaming, the P2P systems have the advantages of lower infrastructure cost and higher capacity. However, the quality of the reconstructed video sequences may become unstable by a direct utilization of P2P network for video streaming. This is because any peer node is allowed to freely leave╱enter the system. Therefore, it may be possible to have an abrupt input sources interruption for the nodes receiving video bitstreams from the leaving peers. Another disadvantage of the P2P system is that many peers may be located in the asymmetric network environments such as ADSL. Since the uplink bandwidth for these peers is usually limited, it may be difficult for them to provide high quality video streaming services. To eliminate the drawbacks stated above, three topics are investigated in this projects: (1) Video encoding╱streaming algorithm well-suited for P2P network, (2) P2P protocol best matched for video streaming, (3) Network simulation tool for evaluating the performance of the proposed algorithm. To design the video coding schemes for P2P networks, layered multiple- description coding technique is adopted. The P2P network protocol should also be revised to accommodate the employment of the layered multiple description coding scheme for video streaming. In order to evaluate the performance of the proposed P2P video streaming system, this project also involves the development of software tools capable of simulating large-scale P2P systems. The tool integrates the description of the P2P protocol into the simulation environment, and allows the evaluation and observation of the video streaming system to be performed in the packet level. We therefore can view the tool as an important platform for the performance measurement of P2P video streaming systems.