UMLsecRT

We present our article on UMLsecRT, a model-based security engineering technique that allows monitoring of Java-based systems for compliance with their security design, that was published as journal-first article in the Transactions on Software Engineering (TSE) [PBJ24a]. Despite a significant focus on security by design in academia and practice, OWASP continues to highlight insecure design as a major threat. Current runtime security monitors are cumbersome to configure and rarely used in practice. UMLsecRT simplifies this process by allowing security experts to annotate design models with security properties, which are automatically synchronized with corresponding source code annotations for efficient run-time monitoring. This technique allows to detect and react to security violations at runtime and facilitates the improvement of a system’s security design based on detected security violations. We evaluated the effectiveness, applicability, and scalability of UMLsecRT in experiments based on the DaCapo benchmark, real-world vulnerabilities, and open-source projects of various sizes. We conducted user studies on usability and practical feasibility. Two case studies, the Eclipse Secure Storage and the iTrust electronic health record system, demonstrate the potential of UMLsecRT to improve security engineering practices.