Increasing trends in pervasive computing demonstrate a requirement for context awareness. The security problem has also become a key issue with context awareness. Access control should execute its decisions by capturing security-relevant context, such as time, location, user activity, and other environmental information available when the access requests arrive. In previous context-aware access control systems, a query issued by an authorized user could not be answered when the context specified in access control policy do not exactly match that specified in the query, even though both are semantically related. In this paper, Semantic Context-aware Access Control (SCAC), is proposed, to solve the problem mentioned prior. The proposed SCAC system takes contexts and its ontologies from context middleware and subsequently arranges contexts according to the abstraction level, to build context hierarchies. Using context hierarchies and reasoning rules extracted from the context ontologies, SCAC can overcome the semantic gap between contexts specified in the policy and contexts collected from the dynamic context sources in pervasive environments.
