Double Level-based P2P Trust Model Research

As a new transmission mode, P2P provides privacy protection and anonymous communication to users, and it makes the network users'communication more and more open and flexible. However, the decentralized nature leads to P2P brings some security risks to network. The existing solutions for security issues are authentication, authorization, and confidence mechanism and so on. As an important component of trust mechanism, trust model is a hot issue in solving the security problems.At present, there are a lot of trust models are not full consideration in the design principles and methods, resulting in trust model scalability and the nodes' security, stability become poor. Therefore, this article through the analysis and research about the existing trust models' shortcomings, mainly complicating the following works:(1) The existing single trust model in P2P network has some shortcomings such as the poor scalability problem, communication bottlenecks and so on. Therefore, this article establishes a double level trust model, the model divides the P2P network into many clusters and the trust relationship is divided into inside and outside of cluster trust.The trust model largely reduces the complexity of computational trust relationships, and improves the network performance and system scalability.(2) The existing cluster selection algorithms has the deviation and affects network stability, this paper presents a cluster head selection algorithm. The algorithm has the candidate mechanism and is improved based on Bully algorithm, it greatly improves the quality of cluster head in system, so that the P2P system will more security and stability.(3) According to the different trust relationships, this paper will divided the credibility calculation method into inside and outside of cluster calculation methods. The method is simple and has the strong reliability.(4) In order to regulate the behavior of the nodes and improve the success rate of transactions, this paper proposes some security mechanisms. For example, restrain the malicious nodes' behavior and encourage the good nodes.These mechanisms will make the model adjust to variation P2P application environments.In the simulation experiment, through the contradistinction about the cluster head seletion algorithm, it verifies that the improved selection algorithm will improves system stability and security; it also verifies that this trust model will control the malicious nodes and improve network performance and so on; after comparing the various models, it verifies this model's effectiveness and feasibility.