论文部分内容阅读
为了改善目前的gossip协议收敛时间较长以及易于产生大量非Internet服务提供商(ISP)友好流量的缺陷,提出一种具有群体感知能力的gossip协议,通过在中心服务器上进行用户聚类分析,使得每个对等(P2P)节点能有效判断自身的相对位置,提高邻居选择的效率,并能在跨运营商的对等环境中有效降低非ISP友好的流量.理论分析和仿真结果表明,所提的gossip协议能进一步提高对物理网络的使用效率,并在提高网络性能的同时吸引和容纳更多的对等网用户.
In order to improve the current gossip protocol convergence time is longer and easy to generate a large number of non-Internet service provider (ISP) friendly traffic flaw, this paper proposes a gossip protocol with group-sensing capability. Through cluster analysis on the central server, Each P2P node can effectively determine its relative position, improve the efficiency of neighbor selection, and effectively reduce the non-ISP friendly traffic in the cross-carrier peer-to-peer environment.Theoretical analysis and simulation results show that, Gossip protocol to further improve the efficiency of the use of physical networks, and to improve network performance at the same time to attract and accommodate more peer-to-peer users.