P2P流媒体系统中层次化网络拓扑推断技术的研究

Network Topology Hierarchy Inference on P2P Media Streaming System

目前基于P2P技术的应用已经远远超过了传统互联网应用,成为占用互联网带宽最多的网络应用.因此,如何减少P2P应用、特别是P2P流媒体播放系统对骨干网带宽的过度消耗,成为P2P技术中一个急需解决的问题.本文提出利用一种基于网络断层扫描的分层推断方法,以将P2P流媒体系统中的流量限制在较小的网络范围内,从而减少跨网流量、同网之间的骨干网带宽消耗,缓解网络拥塞.在OPNET上的仿真表明,该方法能够适应P2P流媒体的高动态性,有效降低跨网流量以及对骨干网的带宽消耗速度.

Currently,P2P-based applications,prevailing over internet,become one of the greatest internet network-bandwidth consumers.So,It＇s indispensable to reduce the network bandwidth consume of P2P application,especially of P2P media streaming.This paper proposes a fast-inferring method based on network tomography in order to limit the traffic in smaller network,to decrease network-cross traffic and backbone network-bandwidth consume internally,and to suppress the congestions.The simulation shows that such technique can adapt the high dynamics of P2P media streaming system,and effectively slow down the network-cross traffic and backbone network-bandwidth consume.