Research On Trustworthiness Evaluation And Selection Of Cloud Services Based On Dynamic Trust Modeling

Cloud services are cloud computing resources and applications deployed on the Internet or cloud computing platform,and users can access the required cloud services at any time.However,due to the virtualization,openness,transparency,dynamic resource changes and security of cloud computing,users face the diversity and complexity of quality of service(Qo S)when evaluating and selecting cloud services.Other challenges include personalized demand preferences,the trustworthiness of cloud services,security,and because cloud providers mayprovide functionality-similar services indifferent QoS,the user's trust crisisfor the cloud service may arise.In order to ensure that the selected services can meet the needs of high-quality,high security services,effective cloud service trustworthiness evaluation and selection methods are in demand.At present,the existing cloud service trust models,which consider the specific system and environment,and takeinto account the time decay,the amount of attenuation,trust chain and reputation,haveachieved some research results.However,with the continuous increasement of user needs and the complexity of the cloud service environment,the following challenges have emerged: the system "cold start" problem,the user for different service levels,recommended trustworthiness accuracy,malicious behavior suppression,and user data storage security.Therefore,there are great theoretical and practical significance to construct a dynamic trust model which can adapt to the cloud computing environment and can effectively evaluate cloud services.The main contents of this dissertation are as follows:(1)In order to solve the problem of "cold start" caused by the poor user's evaluation information,and the user's preference for each service attribute in the direct trust degree,amodel of cloud service trust evaluation based on combination weight and gray relational analysis is proposed.The rough weight attribute calculation method and Analytic Hierarchy Process(AHP)are used to calculate the combined weight of direct trust,which effectively reflects the user's service attribute preference.The gray relational analysis method is used to calculate the similarity of the recommended trust degree,so that the recommended trust degree is more accurate,and the comprehensive trust degree is calculated according to the direct trust,recommended trust and reputation.(2)A dynamic cloud service selection model based on trust and SLA is proposed for users' different service level requirements and direct trust update.The SLA monitoring module manages the SLA level,cloud servcesare divided into different levels according to the cloud service capability,allowing the user to quickly select the appropriate SLA.Firstly,In this model,it considers the choice of effective cloud service evaluation factor which is set in the direct trust calculation,and considers the attenuation factor,thenit evaluates the similarity between the service requester and other users,so the reputation of the cloud service more accurately.Secondly,the direct trust,indirect trust and reputation are given different user preferences weight,and a reasonable calculation of the final comprehensive trust value.Finally,the experiment proves that the selection model has high service satisfaction and interactive success rate.(3)Aiming at the problem of user recommendation trust and node malicious behavior suppression,a cloud service selection method based on trust and user preference clustering is proposed.Firstly,the similarity degree of user service attribute preference is calculated by modifying the cosine similarity method.Then,the user services attribute score and user evaluation similarities are calculated by the cloud model-based method,and the similarity degree between each user is obtained.Secondly,we use the improved hierarchical clustering algorithm to cluster the users to form the user preference domain.Finally,an effective trust recommendation is calculated by the best user preference domain.In the calculation of direct trust,the transaction time,transaction amount,penalty factor are all taken into account.(4)Aiming at the requirements of cloud service data storage security and privacy protection,a cloud service evaluation model based on trust and privacy perception is proposed.The entropy weighting method and the AHP method are used to compute the combined weights of direct trust.This model proposes a privacy data protection method based on normal cloud model,which effectively protects data security.In order to solve the problem of dynamic change with the cloud service capability more effectively,the parameters such as customer satisfaction,decay timeand transaction amount factor are introduced to dynamically update the direct trust degree.