Modeling, analyzing and predicting security cascading attacks in smart buildings systems-of-systems

Abstract Software systems intelligence and complexity have been continuously increasing to deliver more and more features to support business critical and mission critical processes in numerous domains such as defense, health-care, and smart cities. Contemporary software-based solutions are composed of several software systems, that form System-of-Systems (SoS). SoS differentiating characteristics, such as emergent behavior, introduce specific issues that render their security modeling, simulation and analysis a critical challenge. The aim of this work is to investigate how Software Engineering (SE) approaches can be leveraged to model and analyze secure SoS solutions for predicting high impact (cascading) attacks at the architecture stage. In order to achieve this objective, we propose a Model Driven Engineering method, Systems-of-Systems Security (SoSSec), that comprises: (1) a modeling language (SoSSecML) for secure SoS modeling and (2) Multi-Agent Systems (MAS) for security analysis of SoS architectures. To illustrate our proposed approach in terms of modeling, simulating, and discovering attacks, we have conducted a case study on a real-life smart building SoS, the Adelaide University Health and Medical School (AHMS). The results from this case study demonstrate that our proposed method discovers cascading attacks comprising of a number of individual attacks, such as a Denial of Service, that arise from a succession of exploited vulnerabilities through interactions among the constituent systems of SoS. In future work, we intend to extend SoSSec to address diverse unknown emergent behaviors and non-functional properties such as safety and trust.