可分级P2P流媒体的自适应传输调度算法

Self-adaptive Transmission Scheduling Mechanism with Network Coding for Adaptive P2P Streaming

可分级视频编码(SVC)技术实现了从单一码流中得到不同帧率、分辨率和图像质量的视频数据,使其更能应对网络的抖动.P2P技术已广泛应用到流媒体直播系统中,现有的SVC P2P传输调度机制主要基于传统流媒体,较少考虑SVC流媒体多层的特殊结构.本文提出一种基于层间网络编码的SVC P2P传输调度算法,称为可分级P2P流媒体的自适应传输调度算法.该算法通过预测和调整邻居节点各层的发送概率,使得请求节点能够按照预定的各层接收比例接收数据包.理论分析和仿真结果显示该算法具有较目前主流算法更好的性能.

The peer-to-peer technology has been successfully used in live multimedia streaming. The scalable extension of the H. 264/ AVC standard（SVC） video streaming is more scalable to the network fluctuations with the multi-layer structure. The transmission scheduling mechanism of SVC P2P streaming is different with traditional single layer P2P streaming. To fully explore the advantage of SVC and P2P to transmit stream, the scheduling problem is urgent to be solved. Several solutions have been proposed, but these solutions are confined to the traditional methods and without fully considering the characteristics of SVC. In this paper, we present a novel transmission scheduling mechanism based on intra-layer network coding scheme for SVC P2P streaming, called Self-Adaptive Transmission scheduling Mechanism（SATM）. In SATM, through forecasts and adjusts the neighbors sending probability of each lay- er, the request node＇s received packets ratio of each layer can achieve the set proportion. The theoretical analysis and the simulation results show that SATM is feasible and has a good performance.