基于图着色的P2P流媒体数据调度算法

Graph coloring-based scheduling algorithm for P2P media streaming

为了提高P2P流媒体的传输性能,提出一种基于图着色的数据调度算法。算法规定加入系统的每个节点和数据块都被分配某种颜色,请求节点优先获取相同颜色的数据块。根据数据稀有性、紧急性、新鲜性定义数据块优先级别,同时评估伙伴节点数据提供能力,以选择恰当的数据提供者。该算法能有效平衡节点负载,合理利用系统带宽,使数据分布更均匀,提高了系统传输性能。仿真实验表明,算法在数据填充率、启动延迟、数据到达率等重要指标上都优于传统数据调度算法。

／ To increase the transmission performance of P2P media streaming, the Graph Coloring-Based （GCB） scheduling algorithm was proposed. In GCB scheduling algorithm, every peer and every chunk were allocated with one color, and the chunk which had the same color as request peer would be required firstly. In order to select the perfect chunk＇s supplier, the algorithm considered the urgency, rarity and freshness of chunks, and then defined the priority of chunks and evaluated the supply capacity of neighbors. With GCB scheduling algorithm, the load balance of nodes was strengthened, and the bandwidth of system was used efficiently, and data distribution was uniformed; therefore, the transmission performance of system was well improved. The simulation results show that the proposed GCB scheduling algorithm for peer-to-peer media streaming outperforms conventional scheduling algorithms in the data filling rate, the start delay, the data arrival rate and so on.