The Assessment Model Of Clouduser Behavior Based On MEBN

At present, the cloud computing has become a major focus in the IT industry. In the international arena, such as IBM, Microsoft, Google and other domestic companies such as Baidu, Alibaba, SNDA, MI have launched cloud services to the public. Cloud service providers or cloud users will lose a part of control of the data or resources because of the service model of the cloud computing. With the vigorous development of cloud computing, research on the cloud security is one of the hot spots of the domestic and foreign scholars. The core problem to solve cloud security cloud is how to solve the trust. The research of trust contains that user trust, cloud service providers and the data transmission trust over the network. When users access cloud resources and services which provided by the cloud service provider, the research on the cloud user trust will be very important, because that many security threats are usually initiated by the user. Cloud user’s data and business still will be lost occasionally.The trust research of cloud users covers both identity trust and behavior trust. We choose to do research on the cloud user behavior trust aiming on user-generated threats, such as avoidance system security, losses which caused by accounts loss or account theft, analysis the factors which influencing their behavior trust. Considering the multiple perspectives such as the behavior pattern of cloud user and the trust of the environment, we give definitions of multiple properties that affecting cloud user behavior assessment. Using multi-entity Bayesian network model to assess the credibility of the cloud user behaviors. To solve problems such as malicious actions, account loss and other threats to account illegal operation, we take measures to reduce the threat degree from internal users such as dividing the abnormal login time into several sections and considering historical trust and other methods. We divide trust into interval levels and make the assessment results be matched and use the average variance to establish the credibility of the cloud user behavior signatures. In order to test this method to detect abnormal behavior of the internal user effect, we do simulation experiment by using some data and compare the results with several other methods, finally give a comprehensive evaluation of the advantages and disadvantages of this method.