Self-Adaptive Role-Based Access Control for Business Processes.

We present an approach for dynamically reconfiguring the role-based access control (RBAC) of information systems running business processes, to protect them against insider threats. The new approach uses business process execution traces and stochastic model checking to establish confidence intervals for key measurable attributes of user behaviour, and thus to identify and adaptively demote users who misuse their access permissions maliciously or accidentally. We implemented and evaluated the approach and its policy specification formalism for a real IT support business process, showing their ability to express and apply a broad range of self-adaptive RBAC policies.