非结构化P2P网络基于马尔科夫链的搜索算法研究

摘要

Existing resources searching algorithm of unstructured P2P network has not considered the interests of node combining with the load of the whole network.The paper not only considers the node search resources query based on interest,but also considers the load information of each node.Thus,this paper puts forward the resources searching algorithm,BoMC,which is based on Markov Chain model.The BoMC algorithm makes full use of Markov Chain to discrete transition probability matrix for unstructured P2P network though random sampling.However,transi-tion probability is based on forwarding factor of the node that contains of the information of interests and load.On the basis of the characteristics of the stationary distribution of Markov Chain,the algorithm could make the entire network become the convergence condition.And in the process of query,the proposed algorithm would update transition proba-bility dynamically.In PeerSim simulation environment,this paper realizes resources searching algorithm BoMC and discuss the traditional P2P resources searching algorithm with the proposed algorithm.%现有的非结构化 P2P 资源搜索算法并没有将兴趣与负载结合进行考虑,本文不仅考虑节点搜索资源时基于兴趣的查询转发,也综合考虑了各个节点的负载信息.基于此,本文设计提出了基于Markov Chain 模型的资源搜索改进算法BoMC.BoMC算法利用马尔科夫模型为非结构化P2P网络节点通过随机采样建立状态转移概率矩阵.而转移概率是基于节点的转发因子,其中包含有节点兴趣及负载的综合信息.我们知道,基于马尔科夫链平稳分布的特性可以使整个网络在查询过程中趋于收敛状态,进而达到节点的负载均衡.根据网络负载分布情况,该算法考虑到动态更新转移概率.在PeerSim的仿真环境下,实现BoMC算法并将其与传统的P2P资源搜索算法作比较.