Research Of Trust Management Model And Evaluation Methods For Cloud Services

Because of its large scale,virtualization,high reliability,high scalability,pay-on-demand and other advantages,cloud computing provides the users large convenience for complex calculation,bulk storage,etc.However,with the popularity of cloud computing,cloud security issues have been exposed.Due to a large number of terminal equipments access,cloud services cannot judge the authenticity of requesting users.It is also difficult for users to distinguish the credibility of cloud services when they choose cloud services.Therefore,it needs an effective trust management mechanism to ensure trust relationship between cloud services and users.As the traditional certificate-based trust management cannot solve the trust crisis between users and cloud services effectively,the behavior-based trust management is gradually applied in the security issues of cloud services.As the current trust management does not consider users' trust value,the paper puts forward a new trust management model based on two-way evaluation.The paper joins the users' trust evaluation in the trust mechanism,looks for malicious users retroactively by reverse thinking and punishes them.The model prevents malicious users and malicious cloud services by evaluating the trust of both of them.In view of the similarities of behavior between users,the paper proposes an improved EigenTrust trust evaluation algorithm based on similarity.On this basis of EigenTrust,the paper adopts multiple satisfaction evaluation instead of binary satisfaction evaluation to improve the accuracy of satisfaction evaluation of cloud service.The paper adopts cosine similarity to calculate behavior similarity between users to improve the accuracy of indirect trust evaluation.Because the determination of weights in decision algorithm influences the outcome of the decision directly,the paper puts forward an improved TOPSIS multi-attribute decision algorithm based on information entropy.Complying with the objective principle in best service selection,the paper introduces the TOPSIS algorithm.Allowing for the stability of multi-attribute indexes,the paper uses the information entropy as the weights of multi-attributes.The improved algorithm notonly guarantees the objectivity of the weights determination method,but also makes decision results more truthful.Experiments prove that the improved EigenTrust trust evaluation algorithm based on similarity can evaluate the trust value of cloud services more accurately and identify malicious users and cloud services more effectively.The improved TOPSIS multi-attribute decision algorithm based on information entropy can choose the best services for users more objectively.