Safety-Security Analysis via Attack-Fault-Defense Trees: Semantics and Cut Set Metrics

Cyber-physical systems such as the advanced smart grid have a dynamic interaction between security, safety, and defense. Therefore, we need risk management strategies that take all three into account. This paper introduces a novel framework that seamlessly combines attack trees, fault trees, and defense mechanisms: Attack-Fault-Defense Trees (AFDTs). This model creates a common language using an easily understood visual aid designed for experts from various backgrounds, thereby accelerating multidisciplinary collaboration. We define the semantics and cut set metrics for AFDTs and explore how qualitative analysis can be done through cut set analysis. Furthermore, we provide a case study that revolves around a Gridshield lab, which is a set of remotely connected charging stations at the University of Twente campus, to demonstrate the practical implementation of AFDT analysis.