基于RTT优先位图匹配结合的邻居选择P2P拓扑  

Survey on P2P Neighbors' Selection Based on RTT and Bitmap

作　　者：高强[1] 周绪俊 李松江[1] 李艳莉[1] 

机构地区：[1]吉林大学计算机科学与技术学院,长春130012

出　　处：《北京工业大学学报》2014年第9期1418-1421,共4页Journal of Beijing University of Technology

基　　金：国家自然科学基金资助项目(61073164)

摘　　要：为解决随机的拓扑结构P2P流媒体系统中主干网络压力大、启动延迟长和帧丢失率高等问题,提出了一种基于往返时延(RTT)优先位图结合的邻居选择的算法.该算法利用Tracker服务器存储整个网络中所有节点之间的RTT信息和Buffer-Map的位图信息,节点向Tracker服务器发送邻居请求,Tracker服务器根据节点的RTT信息和位图信息及邻居筛选函数为节点生成最佳的邻居列表返回给节点.仿真实验证明:相对于随机的邻居选择算法,该算法有效地降低了节点的启动延迟、播放延迟和帧丢失率.To solve the high backbone network pressure, long startup delay and high frame loss rate of the random P2P streaming system, a neighbors＇ selection algorithm based on round-trip time （RTT） and bitmap was proposed. The algorithm used the Tracker to store the round-trip delay between nodes and the Buffer-Map bitmap across the network, and the node sent a request to the Tracker, then the Tracker sent the best neighbor list based on the node＇s RTT information, bitmap information and the screening function to the node. The simulation result shows that the algorithm effectively reduces the node startup delay, playback delay and frame loss rate compared with a random neighbors＇ selection algorithm.

关 键 词：邻居选择 位图信息 往返时延(RTT) 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]