P2P流媒体直播系统数据调度算法及仿真

Data Scheduling Algorithm and Simulation in P2P Live Media Streaming

在数据驱动的P2P流媒体直播系统的研究中,数据调度算法的优劣影响流媒体的播放质量。因此主要研究了P2P流媒体直播系统中的数据调度问题。通过定义请求数据块的播放质量优先级,提出了最大化播放质量优先级的分布式调度模型。模型首先预测出本周期内请求节点与邻居节点之间的实际带宽,然后建立最优化数学模型并将其转换为等价的指派问题,根据该指派问题构造等价的赋权完全二部图,最后利用Kuhn-Munkres算法求出本周期的数据块调度策略。利用P2PStrm Sim仿真器仿真,结果表明,所提出的分布式调度算法的性能比传统调度策略有显著提高。

Data scheduling algorithm affects the playback quality of media streaming in data - driven P2P live streaming system. So,the data scheduling problem in the P2P live streaming system was studied in this paper. By the definition of playback quality priority of data block, this paper presented a distributed scheduling model to maximize the playback quality priority. For this model, the actual bandwidths were calculated which were between request node and its neighbors in this cycle. Then, the optimal mathematic model was presented and transformed into an equivalent assignment problem. Finally, according to this assignment problem, a weighted complete bipartite graph was construc- ted, and a strategy of data block scheduling in this cycle was given by using Kuhn - Munkres algorithm. By P2p Strm- Sire,the simulation results show that the performance of the proposed distributed algorithm is superior to some tradi- tional data block scheduling strategies.