P2P环境下基于贝叶斯理论的信任模型研究

发布时间：2018-03-14 04:48

本文选题：对等网络　切入点：Beta信任模型　出处：《安徽师范大学》2014年硕士论文　论文类型：学位论文

【摘要】：对等网络(简称P2P网络)在分布式存储、协同工作、大规模并行计算等方面表现出巨大的优势，成为新的研究热点。然而，P2P具有的开放性、匿名性以及节点之间松耦合等特点使得节点可以肆意散布非法内容，滥用网络资源，，导致节点之间信任缺失，严重阻碍了P2P的进一步发展。信任反映的是一个主体对另一个主体行为以及能力的综合评价，P2P网络中节点之间的信任是促使节点合作、激励节点共享资源的有力保障，直接影响到P2P系统的整体性能。本文主要围绕P2P环境下基于贝叶斯理论的信任模型进行研究，主要工作如下： (1)应用贝叶斯理论中的Beta分布计算信任度的方法，提出了一种非结构化P2P环境下的Beta信任模型，该模型通过自适应算法构建非结构化网络拓扑，能够对信任度进行计算、存储以及动态更新，实现了完整的信任模型体系结构。 (2)应用贝叶斯理论的Dirichlet分布计算信任度的方法，提出了一种结构化P2P环境下的Dirichlet双信任模型，将信任分为服务信任和反馈信任，用于度量节点的不同行为，从而提高信任计算的准确度。 (3)针对P2P环境中常见的安全问题，如搭便车(Free Riding)、诋毁(Slander)、共谋(Collusion)、波动攻击(Oscillation Attack)、策略攻击(Strategy Attack)等，基于以上恶意行为特有的行为属性，提出了相应的防护机制。仿真实验表明，本文信任模型相对于其他信任模型在抵御这些攻击方面表现出更好的性能。
[Abstract]:Peer-to-Peer (P2P) network has great advantages in distributed storage, collaborative work and large-scale parallel computing, and has become a new research hotspot. Anonymity and loose coupling between nodes enable nodes to wantonly distribute illegal content abuse network resources and lead to lack of trust between nodes which seriously hinders the further development of P2P. Trust reflects the comprehensive evaluation of the behavior and ability of one entity to another. The trust between nodes in P2P network is a powerful guarantee to encourage nodes to cooperate and encourage nodes to share resources. This paper focuses on the trust model based on Bayesian theory in P2P environment. The main work is as follows:. 1) using the Beta distribution in Bayesian theory to calculate the trust degree, a Beta trust model in unstructured P2P environment is proposed. The model can calculate the trust degree by using adaptive algorithm to construct the unstructured network topology. Storage and dynamic update to achieve a complete trust model architecture. 2) applying the Dirichlet distribution of Bayesian theory to calculate the trust degree, a Dirichlet double trust model in structured P2P environment is proposed, which divides trust into service trust and feedback trust, which is used to measure the different behaviors of nodes. In order to improve the accuracy of trust calculation. 3) aiming at the common security problems in P2P environment, such as free rider, slander, collusion, Oscillation attack, strategy attack and so on, the corresponding protection mechanism is put forward based on the specific behavior attributes of the above malicious acts. The simulation results show that, Compared with other trust models, this trust model has better performance in resisting these attacks.
【学位授予单位】：安徽师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.08;TP393.02

【相似文献】