Research On User Behavior Trust Model For Personalized Cloud Computing

With the growing demand of users, as a new person-centered calculation model,cloud computing becomes the research focus of IT area. At present, how to establishtrust model which suits the environment of cloud computing has been one of the hotspots and difficulties of its security area in the academic world. In order to enhancethe accuracy of trust evaluation between users and service providers, this thesisdiscusses the method of how to establish a fair and objective user behavior trustmodel. The main contributions are listed as follows:1) To evaluate the trust relationship between users and service providerscompletely and consider the uncertainty and the dynamic of service resources, adynamic trust model for personalized cloud services (PCSDTM) is put forward. Therelevant knowledge of grey system theory is used to reflect the subjectivecharacteristics of trust; User behavior incentive system and punishment mechanismare constructed to give the network a sound progress; A new dynamic distributionmethod of direct trust and indirect trust factor is proposed based on evaluationsimilarity. Simulation results show that PCSDTM is more accuracy and scientific.2) Integral security of information interaction between users and serviceproviders will be analyzed deeply in the environment of cloud computing. Therefore,the P3Trust model based on user type and privacy protection for personalized cloudservices will be constructed. Firstly, the trust evaluation agency and public keycryptography is introduced and then a trust evaluation system framework based onprivacy preserving is designed to control conspiracy cheat, malicious recommendationthrough protecting interactive sides’ privacy, which makes the results of the trustmuch more real; Secondly, the trust evaluation method based on user type is putforward to improve the veracity and efficiency of trust assessment by adjusting therelevant factors of trust completely; A new trust update mechanism based on qualityof service is introduced to depict the trust dynamic. Lastly, simulation results showthat P3Trust can evaluate the trust relationship effectively and withstand harshenvironments by protecting users’ privacy.