Research On Network Coding Design In Heterogeneous P2P Streaming

Providing high-quality streaming over Peer-to-Peer (P2P) systems faces multiple challenges due to high heterogeneity of receivers in terms of access bandwidths, buffering levels, initial playback delays, and so on. Network coding has been shown to improve the throughput and bring various performance benefits in the P2P streaming systems. In order to satisfy diverse playback quality of video in heterogeneous P2P streaming systems and simultaneously improve the system performance, we study the heterogeneous P2P streaming based on network coding. The main contribution of this thesis is listed as following:1. Based on random network coding, a network coding segmentation design scheme for P2P streaming is proposed in this paper. The network coding segmentation design scheme is based on bandwidth smoothing model to establish feasibility discriminate module of network coding segmentation. Then, we compute the maximum network coding segmentation under peer configuration resources, as well as the network coding segmentation constraint of each peer. Due to different peers with different segmentation constraints, peers are grades into different ranks based on the segmentation constraints. Peers with different ranks can operate cooperatively with each other. The low-level peers can employ the encoding data of the high-level peers.2. A network coding and data schedule method on the basis of fine granularity scalability (FGS) for Peer-to-Peer streaming is proposed in the paper. We first design a cross-level network coding scheme that independently divides video stream into small-size segments according to the belonging priority, and then perform network coding on multiple levels of segments with the same timeline. Consequently, we can achieve perfect coordination among peers even though they adopt different quality levels. Besides this scheme, a rank based data schedule to smoothly switch quality levels is proposed, which is implemented in each node proposed scheme can obtain good playback smoothness with continuity index close to 1.0. Simulation results demonstrate that our proposed scheme can obtain good playback smoothness with continuity index close to 1.0, and outperforms traditional data schedules, especially in the case of low bandwidth, small-size buffer, and short playback delay.In this paper, we study on network coding design in heterogeneous P2P streaming. The above achieved results can provide basis for the design of the network coding scheme and network coding scheduling algorithms, and further promote the development of practical network coding.