The Research of Trust Model and Related Technologies in P2P Networks
|School||Nanjing University of Posts and Telecommunications|
|Course||Applied Computer Technology|
|Keywords||P2P gray forecasting Cloud model Simulated Annealing trust model|
The application based on P2P (Peer-to-Peer) technology is one of the most important applications of the Internet. P2P networks, with the features of file-sharing and no centralized servers, have caused wide public concern. However, a P2P system is large in scale, and the component nodes are dynamic, so security issues are very serious in it. The research of security issues in P2P networks has become one of the hottest topics in academic circles.This thesis focuses on the researches of free-riding behavior, trust model and search algorithms of trust information in P2P networks. The main achievements are as follows.A free-rider number forecasting model based on Gray System Theory—GST model is constructed. Taking into account that severe free-riding behavior will constrain the healthy development of P2P networks, and using a certain method, Gray System Theory is able to change a messy data sequence into a regular data sequence, which can be used for forecasting the number of free-riders in P2P networks. Experimental results show that GST model is capable of making reasonable predictions on the number of free-riders in normal P2P networks.A Cloud-based weighted trust model—CBWT model is proposed. This model consists of two algorithms: Weighted Trust Information Transfer Algorithm and Weighted Trust Information Combination Algorithm. Experimental results show that this model is capable of calculating the trust degrees between two nodes in P2P networks and it provides a reference for the trust decision of the node in P2P networks.A Simulated Annealing based search model is designed. Taking into account the defects of the search algorithms in existing P2P networks and the advantages of Simulated Annealing in solving optimization problems, apply Simulated Annealing into the searching of nodes’ trust information in P2P networks. Experimental results demonstrate that this model is capable of searching or updating nodes’ trust information in all kinds of P2P networks.Through the research of trust model and related technologies in P2P networks, the method of forecasting the number of free-riders is constructed, the strategy of the transferring and combining of the trust information is proposed, and the algorithm of searching the trust information is also designed. The work in this thesis is not only providing a reference for the research of security issues in P2P networks, but also can be integrated into existing trust model of P2P networks for the purpose of promoting the overall performance of the existing model and the security of the entire P2P networks.