Making CPS Modeling Easier and Reusable – University, NIST Researchers Propose a Way

Large cyber-physical systems (CPS), in which operational technology systems monitor and control the physical environment, such as through services deployed for smart cities, typically consist of many independent subsystems that must interoperate. These CPS, or equivalently, Internet of Things (IoT) systems, rely on communications networks to relay timely messages across the system, thus allowing sensing and responding to changing situations. Ensuring such CPS work as intended requires accurately simulating their performance; this involves integrating various models associated with CPS subsystems, or “co-simulation.” Yet, such CPS modeling has been challenging due to network simulation complexity and customization needed for specific domains.

Vanderbilt University and NIST researchers offered a novel approach to meeting this need in their paper,  Reusable Network Simulation for CPS Co-Simulations, presented at the CPS-IoT Week '23. The researchers propose rapidly synthesizing large-scale, integrated simulations using a model-based framework. The framework transforms models and their information exchanges into code that complies in alignment with IEEE’s High-Level Architecture, which enables the models to work together as something called a co-simulation. The framework implementation also includes a reusable network component, which allows the co-simulation to be configured for use in different domains and scenarios.

The researchers also updated their reusable cyber-attack library, which they integrated with the co-simulation to allow testing of a CPS under various attack scenarios. The paper includes a case study which involves simulating the effects of attacks on the networking infrastructure of smart cities. Specifically, in this case study, malicious actors attack traffic light operations with the city’s command and control conducting counter operations.

In the future, researchers plan to focus on multiple network simulations, scalability of simulation, and adding cyber-defense to the cyberattack library, thus allowing experimentation with cyber-gaming scenarios.