P2P Streaming Based On Scalable Video Coding

With the rapid development of Internet technology, people can access the Internet easily to retrieve information. Internet video service as a multimedia application is becoming more popular. According to statistics, the video services have occupied the about 40% of current Internet traffic and still keep growing. Cisco published a report to indicate that in 2013 network video traffic will be up to 90% of the total traffic. It is obvious that streaming applications have broad prospects. But with the increase of the service scale, user's requirement of quality of service (QoS) will be improved. How to provide high quality, low load and scalable streaming service is still challenging.Video coding technology is a key factor of the quality of service and it has received widespread attention of researchers. H.264/AVC based scalable video coding technology which can provide scalability to adapt to complex, heterogeneous network environment gives us a promising solution for streaming media services. Scalable video coding supports temporal scalability, spatial scalability and quality scalability. Scalable video streaming can provide scalable video service in heterogeneous network environment and various capabilities of end-user devices. But the scalability needs more computation which reduces the coding efficiency. Therefore, there are still lots of improvement to be done when introducing SVC to real world applications.In the large-scale streaming service, how to utilize the bandwidth resources in the network to reduce pressure on server load is another key problem. The peer to peer (P2P) network could utilize the bandwidth of peers in the network to forward video clips which can efficiently reduce the server load. How to design a suitable P2P protocol for SVC streaming service is an important research area.In this thesis, we proposed a SVC based P2P streaming system to address the above issues. For video coding, by modifying some algorithms and parallelizing the coding process an improved SVC encoder based on JSVM has been proposed. For video transmission, a BT based P2P protocol has been designed by modifying the scheduling algorithm and incentive mechanism to combine with the feature of SVC streaming.