A non-intrusive runtime enforcement on behaviors of open supervisory control and data acquisition systems

Abstract

With the development of Web-based services and related technologies, today’s supervisory control and data acquisition is regarded as an Internet of things service system for industrial infrastructures and ensuring open supervisory control and data acquisition systems in a secure and safe state at runtime becomes a critical and mandatory requirement. Existing host-based monitoring automata are vulnerable because “inside” malware may compromise and subvert the monitoring mechanism itself, and the virtual machine–based monitoring cannot provide observable running traces of the protected services because of the isolation between these services and the runtime monitor. In this article, we propose a non-intrusive solution to guarantee runtime state of open supervisory control and data acquisition systems. In this solution, the running traces of protected services are obtained in an “out-of-box” framework, which is built on abstract execution of network events on Internet of things service models and virtual machine semantic reconstruction of the protected services. In addition, a property checking procedure is employed to check the states of physical devices in advance to guarantee the runtime behavior in compliance with the security policies of open supervisory control and data acquisition systems. In such a way, the solution provides a fine-grained protection for open supervisory control and data acquisition systems and physical devices running in safety.