Font Size: a A A

Research On P2P Trust Model And Search Algorithm Based On The Incentive Mechanism

Posted on:2012-12-05Degree:MasterType:Thesis
Country:ChinaCandidate:F LiFull Text:PDF
GTID:2218330344450795Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In recent years, Peer-to-Peer is very hot in computer network. Its advantages, such as, fault-tolerance, expandability of resources sharing, self-organizing, load balancing, anonymity, developed rapidly in last several years and the application of Peer-to-Peer network became more and more extensive. However, resource search in P2P network is always a very important issue.Resource search is classified into structured and unstructured resource search according to different P2P network structure. This paper probes into the unstructured P2P search algorithm. The common trust model is firstly detailed analyzed and improved scheme is provided based on the common trust model. Secondly, an improved scheme on random walk search algorithm is put forward also based on the trust model. The research and results of this thesis are as following:1. The research of trust model based on the feedback credibility. The performance of P2P network searching is greatly depend on the cooperative forwarding and sharing among peers, however, most P2P clients are selfish because forwarding and other clients'query will consume their own resources. Therefore, they only download from P2P and never contribute to P2P network. Since the existence of selfish peers has great impact on searching efficiency, to distinguish the selfish peers, we need to build a high-efficient, scalable and secure trust model. In this thesis, a trust model based on the feedback credibility is put forward. The trust model assesses the local credibility of the feedback of other peers to avoid the attack from malicious peers. Simulation tests and analysis results show that the model can distinguish selfish peers.2. The research on the trust model based on the search algorithm. Improved search algorithm of trust model proposes a two-way choice based on the incentive mechanism. When a node is requesting a service node, choose a high-credibility node instead of random walk. In order to avoid the overheating of high-credibility peers and for the benefits of high-credibility peers, service peers distinguish the credibility of peers upon request and provide priority to the high-credibility peers...
Keywords/Search Tags:Unstructured P2P Network, Random Walk, Trust Model, Incentive Mechanism, Resource Search
PDF Full Text Request
Related items