基于马尔科夫链的非结构化P2P网络信誉模型  

Reputation estimation model for unstructured P2P network based on Markov chain

作　　者：贺超凯[1] 吴蒙[2] 

机构地区：[1]南京邮电大学计算机学院,江苏南京210023 [2]南京邮电大学研究生院,江苏南京210003

出　　处：《南京邮电大学学报（自然科学版）》2017年第2期74-79,共6页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

摘　　要：P2P网络节点间建立信任关系是其安全面临的主要问题,信誉模型通过收集节点对交易的信息反馈给节点赋予信誉值。非结构化P2P网络缺少快速散列和搜索机制,如何进行有效的信誉计算是P2P网络信誉模型的挑战。文中提出一种基于马尔科夫链的非结构化P2P网络信誉模型,根据节点在组内的交易计算一跳节点的信誉值,并证明模型是一个遍历连续马尔科夫链模型,一个组内最高信誉值节点将被选为认证节点,次高信誉值节点作为备份认证节点,数值实验结果证明模型的有效性。Building trust relationships between peers is an important and difficult problem for security needs of P2P network without a central server. P2P reputation model is introduced by collecting the generated peer feedbacks and aggregating them to yield global reputation scores. Most P2P applications in the Internet are unstructured, without fast hashing and searching mechanisms, how to perform efficient reputation estimation is a major challenge on unstructured P2P computing. This paper proposes a two-step reputation estimation approach for the unstructured P2P network. Firstly, a Markov chain model is proposed to determine the reputation value for each one-hop neighbors. A peer＇ s reputation value is analyzed from its previous trust manner in this group. The proposed trust model is proven as an ergodic con- tinuous-time Markov chain model. Secondly, a peer with the highest RV of a group is selected as the central authentication server, the peer with the second highest reputation value is selected as the backup group leader taking over central authentication server when the older server fails. The procedures of the peer＇s reputation value are detailed. Numerical results indicate that the analytical reputation value of each peer is very close to that of simulation under various situations.

关 键 词：离散马尔科夫链 信誉计算 非结构化P2P网络 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]