一种基于有序偏好的P2P网络信任模型

发布时间：2018-10-24 10:26

【摘要】：自P2P网络问世以来,就受到了越来越多的关注,现已被使用在各种各样的网络应用中。P2P网络改变了传统网络中的客户端/服务器模式,采用端对端原则,网络中所有实体之间都是平等的,每个实体既是服务的提供者也是服务的享用者。然而,P2P网络的开放性特点也给其安全性带来了影响,阻碍了P2P网络的进一步发展。目前,基于P2P网络的文件共享、视频点播、即时通信、电子商务等应用逐渐成为一种主流应用模式。但在P2P网络中存在一些恶意的参与者,他们通过网络传播病毒,散播广告等。为了解决这一问题,学者们将人类社会关系中信任的概念引入到P2P网络中,以模仿人与人之间的信任关系来建立网络中节点与节点之间的信任关系。现有的研究表明,信任模型可以在一定程度上解决P2P网络中的信任问题,遏制节点的恶意行为。然而,在现有模型中,多采用连续或离散的单一值刻画信任,并依据该信任值做出信任决策,方法过于简单。本文针对现有信任模型中存在的信任刻画粗糙及信任决策方法过于简单的问题,提出一种基于有序偏好的P2P网络信任模型。首先,将服务刻画成n维属性向量,以便能够更加细致地表现节点在各个属性方面的能力；其次,考虑到每个节点对所请求服务的各属性可能存在不同偏好这一事实,我们利用AHP(Analytical Hierarchy Process)方法,将定性化的用户偏好信息转化为定量化的有序偏好,并据此求出请求节点的期望向量。最后,将请求节点的期望向量与每个响应节点的信任向量进行结合,以计算出各响应节点的评估值,以帮助请求节点进行信任决策。模拟实验结果表明,本文的模型能够有效提高节点信任决策的满意度,并能够有效抵抗节点的恶意攻击行为。
[Abstract]:Since the advent of P2P network, more and more attention has been paid to it, which has been used in various network applications. P2P network has changed the client / server mode in traditional network and adopted the end-to-end principle. All entities in the network are equal, and each entity is both a service provider and a service user. However, the open characteristic of P2P network also affects its security and hinders the further development of P2P network. At present, P2P network based file sharing, video on demand, instant messaging, e-commerce and other applications have gradually become a mainstream application mode. However, there are some malicious participants in P2P networks, who spread viruses and advertisements through the network. In order to solve this problem, scholars introduced the concept of trust in human social relations into P2P networks, and established the trust relationship between nodes and nodes in the network by imitating the trust relationship between people. The existing research shows that the trust model can solve the trust problem in P2P network to some extent and restrain the malicious behavior of the nodes. However, in the existing models, it is too simple to use a single continuous or discrete value to describe trust and make a trust decision based on the trust value. In this paper, a P2P trust model based on ordered preference is proposed to solve the problems of rough description of trust and too simple method of trust decision in existing trust models. First, the service is characterized as an n-dimensional attribute vector, so that the ability of the node in each attribute can be represented in greater detail; secondly, considering the fact that each node may have different preferences for each attribute of the requested service, By using the AHP (Analytical Hierarchy Process) method, the qualitative user preference information is transformed into the quantitative ordered preference, and the expectation vector of the request node is obtained. Finally, the expectation vector of the request node is combined with the trust vector of each response node to calculate the evaluation value of each response node to help the requesting node to make the trust decision. The simulation results show that the proposed model can effectively improve the degree of satisfaction of the node trust decision and can effectively resist the malicious attack behavior of the node.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.08

【参考文献】