Using a Novel Method for Trust Evaluation to Enhance ABAC Capabilities

Abstract

Access control is a security mechanism that prevents unauthorized access to sensitive resources. Attribute-Based Access Control model (ABAC) makes decisions by the considerations of subjects’ attributes. Although it has many advantages, it is not dynamic. In dynamic environments, the system should be able to change the users’ permissions according to their manner of activities. So, this paper proposes using trust besides ABAC. The introduced method for the evaluation of trust employs both the Fuzzy Inference System (FIS) and Neural Networks (NN), which is called Fuzzy-Neural based trust (FNT). As trust is evaluated according to some predefined parameters, the proposed model uses FIS to assess each parameter. Next, the assessed parameters should be mixed to generate a single result. Since the definition of an exact function might be difficult and complicated, the proposed model employs the NN, which acts as a black box and generates an expected output after its learning process. For the evaluation of trust, the assessed parameters are fed to the NN to produce a final result. Whenever a subject’s trust is evaluated, then the proposed model makes the final AC decision by the consideration of both ABAC’s result and the amount of trust. Afterwards, we evaluate the proposed model and then highlight its advantages by comparing with some other famous AC models.