考虑节点服务优先级的P2P网络群发算法研究

摘要

In order to promote fairness of node service request in P2P multicast,and solve communication efficiency problem of P2P multicast under the multi-link condition,this paper proposed an integrated optimization communication tree algorithm (IOT).This algorithm based on multi-link concurrent communication model,introduced the concept of node service request queue,comprehensively considered node communication link number and node service priority,and selected several nodes with big communication link number as the communication sources by minor adjustment of node service request queue in order to improve the concurrency and the communication efficiency.Simulation results indicate that IOT algorithm obtains better communication performance than LFT algorithm and QFT algorithm in concurrent communication time,average end-to-end delay,out-of-order rate of queue,node utilization rate and concurrency.Multicast scheme based on IOT algorithm not only can take into consideration the fairness of most nodes' service request priority,but also can effectively shorten the communication time in P2P multicast.%为了提升P2P网络群发中节点服务请求的公平性和解决节点多连接情况下P2P网络群发通信效率问题,提出一种综合优化通信树算法(IOT).该算法在多连接并发通信模型的基础上,引入节点服务请求队列的概念,将节点通信连接数和节点服务优先级综合考虑,通过对服务请求队列局部调整,可将节点通信连接数大的节点优先加入通信树上层,使其在下次通信中作为通信源,进而提高通信并发度和通信效率.模拟实验表明,与连接数优先通信树算法(LFT)、队列优先通信树算法(QFT)相比,综合优化通信树算法在并发通信时间、平均端到端时延、队列乱序率、节点使用率、并发度方面均能取得较好的通信效果.该算法给出的群发方案既能兼顾绝大多数的节点服务请求优先级公平性,又能有效缩短P2P网络群发通信时间.