Efficient Enforcement of Security Policies in Distributed Systems
Policy-based management (PBM) is an adaptable security policy mechanism in information systems (IS) that confirm only authorised users can access resources. A few decades ago, the traditional PBM has focused on closed systems, where enforcement mechanisms are trusted by system administrators who define access control policies. Most of current work on the PBM systems focuses on designing a centralised policy decision point (PDP), the component that evaluates an access request against a policy and reports the decision back, which can have performance and resilience drawbacks. Performance and resilience are a major concern for applications in military, health and national security domains where the performance is desirable to increase situational awareness through collaboration and to decrease the length of the decision making cycle. The centralised PDP also represents a single point of failure. In case of the failure of the centralised PDP, all resources in the system may cease to function. The efficient distribution of enforcement mechanisms is therefore key in building large scale policy managed distributed systems. Moving from the traditional PBM systems to dynamic PBM systems supports dynamic adaptability of behaviour by changing policy without recoding or stopping the system. The SANTA history-based dynamic PBM system has a formal underpinning in Interval Temporal Logic (ITL) allowing for formal analysis and verification to take place. The main aim of the research to automatically distribute enforcement mechanisms in the distributed system in order to provide resilience against network failure whilst preserving efficiency of policy decision making. The policy formalisation is based on SANTA policy model to provide a high level of assurance. The contribution of this work addresses the challenge of performance, manageability and security, by designing a Decentralised PBM framework and a corresponding Distributed Enforcements Architecture (DENAR). The ability of enforcing static and dynamic security policies in DENAR is the prime research issue, which balances the desire to distribute systems for flexibility whilst maintaining sufficient security over operations. Our research developed mechanisms to improve the efficiency of the enforcement of security policy mechanisms and their resilience against network failures in distributed information systems.
- PhD