一种基于层次分析法的超级节点选取机制

在非结构化P2P中,为了使得普通节点能够综合考虑各方面的因素,并根据自身要求选择一个合适的超级节点,提出了基于层次分析法(AHP)的超级节点选取机制SNSM.该机制综合考虑了网络传输往返时延(RTT)、节点兴趣向量相似性以及节点带宽三方面的性能,并在区域组成、资源搜索平均路径长度和网络带宽消耗3个方面和传统的超级节点的随机选取机制进行了比较.仿真实验表明,和随机选取机制相比,SNSM能够在一定程度上提高网络的性能.SNSM能将物理位置上较为靠近的、兴趣较为相似的节点组成了一个区域;不仅减少了25%的信息检索平均路径长度,同时还降低了11%的网络消耗.

AHP-based super node selecting mechanism

In order to conduct the ordinary node to select a super node,a super node selecting mechanism is proposed.The mechanism is named SNSM(SN selecting mechanism),and it is based on AHP(analytic hierarchy process).In SNSM,the ordinary node is allowed to select SN(super node) in its own demand and able to consider multiple factors.Three factors,RTT(round-trip time),IV(interest vector) similarity and bandwidth of the node,are considered synthetically in SNSM.SNSM has been compared to traditional random super node selecting mechanism in three aspects.They are region composition,average path length and network usage.Simulation result shows that the network performance is improved in some degree in SNSM when compared with random selecting mechanism.Based on SNSM,nodes with closer distance and interest are made a region.And average path length is decreased by 25%.At the same time,usage of the network is decreased by 11%.