Policy Based Trust Model in Cloud Computing

Abstract

Cloud computing is a new and promising technology in which cloud resources are provided as a service to its user over a virtualized network like Internet. Cloud computing provides computing and storages resources to its user specially small and medium sized organizations. By providing ”pay per usage” cost model, cloud computing can also reduce cost for use of cloud resources. The issue of trust is one of the major barrier for the usages of cloud computing resources. Nowadays security of the cloud resources is the one of the main factor in cloud computing. Trust in cloud computing can be categorized for the many issue such as access control, prevention, ownership and data security. In this thesis we have label the issue related to user trust and proposed a trust evaluation model for establishing trust based on different policies. In cloud computing, User can directly access and use the cloud resources. The user can perform malicious activities. So it is important to evaluate the user trust based on action, performed by user using the trust model. So to evaluate the user trust, policy based trust evaluation model is used. Based on the user trust value, number of policies breach and the number of warning received the user is classified into different categories. The trust evaluation model provides information to Cloud Service Provider(CSP) to monitor the malicious behavior of user. Based on the trust value of user, the cloud service provider can limit the user from performing activity on the cloud resources and blacklist the user. In future, the number of new user are going to be increase on a large scale so the trust value evaluation for each user will become an overhead. To solve this problem, we have proposed a trust value prediction method. By using this method, we are able to predict the new user trust value with minimum effort. When a new user usages the cloud services, this prediction method predict his trust value based on the previous user performed activities. After predicting the trust value of user, user will be categorized into one of the category set by the CSP.