부산대학교 도서관

Authoring access control policies is challenging and prone to misconfigurations. Access control policies must be conflict-free. Hence, administrators should identify discrepancies between policy specifications and their intended function to avoid violating security principles. This paper aims to demonstrate how to formally verify access control policies. Model checking is used to verify access control properties against policies supported by an access control model. The authors consider Google's Cloud Identity and Access Management (IAM) as a case study and follow NIST's guidelines to verify access control policies automatically. Automated verification using model checking can serve as a valuable tool and assist administrators in assessing the correctness of access control policies. This enables checking violations against security principles and performing security assessments of policies for compliance purposes. The authors demonstrate how to define Google's IAM underlying role-based access control (RBAC) model, specify its supported policies, and formally verify a set of properties through three examples.