【作者】 富志伟；

【导师】 雷建云；

【作者基本信息】 中南民族大学 ， 计算机应用技术， 2011， 硕士

【摘要】 近年来,P2P技术成为人们研究的热点问题,基于其分布式和能充分利用资源的特点快速发展并得到广泛应用。但是,随之产生的安全、效率等方面的问题逐渐凸现,特别是网络中节点的行为缺乏指导和约束,成为限制P2P应用发展的一个重要因素。因此,信任机制在P2P网络中节点信任值的确定及信任关系的传递等问题中的作用尤为突出。而各节点信任值的确定成为信任机制研究的重点。所以信任评估算法的研究为P2P网络的安全性及有效性提供保障。P2P网络技术主要应用在文件共享、协同工作、对等计算、搜索引擎,但文件共享是目前最重要的一个应用。它允许任意终端用户即对等节点通过网络完成文件交换,是一种通过直接交换共享计算机资源和服务的网络体系结构。随着使用P2P网络共享文件的用户越来越多,P2P系统的规模也越来越大。此系统的一个主要挑战就是如何在没有可信第三方或授权中心的情况下确认不同节点之间的信任关系。此时,必须依靠节点过去的行为来决定它是否值得信任,即需要建立信任管理机制。P2P网络中信任机制主要完成获取、分析网络中节点的信任值信息以指导节点间的协作决策;给出正确的信任评估以防止各种欺诈行为;实现与可信节点的协作,并隔离不可信节点等工作。在P2P信任机制方面的研究主要集中在提出合理的信任评估方案和信任度计算方法两个方面,只有尽可能计算出节点的信任度才能为节点之间的安全互操作做出保障进一步保证P2P网络的良好发展。本文总结已有的信任机制研究成果,首先分析了影响信任值的主要因素,提出基于推荐理论获取信任评估的信任管理技术。基于模糊理论的语言变量和模糊推理建立模型。用直接信任值和推荐信任值对P2P网络中节点进行信任评估。根据节点资源类型不同分成不同的兴趣节点群,每个兴趣节点群由一个中心节点来负责管理本群节点的添加和退出。此信任机制通过对信任评估算法的改进增强了P2P系统的安全性。并增加了兴趣节点群的分类并通过TTL(Time To Live)的减值来控制搜索查询的深度,减小搜索范围提高搜索效率,提高了P2P系统的有效性。更多还原

【Abstract】 In recent years, P2P technology becomes a hot issue which gets more and more people’s attention. Based on its distributed feature and the ability to make full use of resources, it has been widely used and rapidly developed. However, the issues of safety and efficiency have gradually appeared, especially lacking guidance and constraint to the behavior of the nodes in the network, which become an important factor in limiting the development of p2p.Therefore, the application of the trust mechanism in P2P network is particularly important in the problems such as the trust value defined and the transfer of trust relation. And the determination of the trust value for each peer becomes the research emphases in trust mechanism. So the research for trust evaluation algorithm provides safeguard for the safety and efficacy of P2P network.P2P network technology is mainly applied in file sharing, collaborative work, equivalence calculation, and search engine. But file sharing is currently one of the most important applications. It allows any end-user namely peer node to complete file exchanging through the network, it is a network system structure that directly exchange resources and services in Shared computers. With the users who use the P2P network sharing files more and more, P2P systems are becoming ever larger. A major challenge of this system is how to confirm a trust relationship between different peers under the condition that without a trusted third party or an authorized center. Meanwhile, the peer must rely on its past behaviors to determine whether it is worthy trusting or not, namely must need to build trust management mechanism.The trust mechanism in P2P network mainly completes the acquisition and analysis of the trust value of nodes in the network to guide decision-making collaboration between nodes, to give the correct trust evaluation in order to prevent all kinds of fraud, to achieve collaboration between nodes, to isolate the untrustworthy nodes. The research of trust mechanism in P2P has focused on proposing a reasonable solution and a algorithm for trust value. Only when the trust value can be calculated as accurately as possible for the node, can it guarantee the security for the interoperability between nodes and ensure the further development of P2P.This thesis summarizes the research results of the existing trust mechanisms. Firstly, analyzes the main factors that influence the trust value and then propose a trust management technology based on the theory of recommended obtaining trust evaluation. Build model based on the theory of fuzzy language variables and fuzzy reasoning. Make trust evaluation for the nodes in P2P network by the direct trust value and recommendation trust value. It divided the network into many groups,according to standard of users’interests.There is a central node to be responsible for the management of the group.This trust mechanism based on the improved trust evaluating algorithms has strengthened the security of P2P system. And by increasing the classification of interest node group and add TTL to control the depth of the search query. Reduce the search scope to improve search efficiency and improve the effectiveness of P2P systems.更多还原