Security Verification of Industrial Control Systems using Partial Model Checking

Abstract

Industrial control systems are moving from isolated to distributed and cloud-connected architectures. While the operational benefits of this migration form the driving force for this trend, the necessary security assurance is often difficult to achieve. Formal methods, including model checking, provide capable technologies to deal with this challenge. However, when formal verification must account for the complexity of modern control systems the state space being explored grows drastically as more details are included in the analysis. This may eventually cause a state space explosion, which makes formal verification infeasible. To address this, we propose a method for decomposing cloud-connected control systems into modules representing the different parts of the system (clients, the cloud, the control network, etc.). Based on the decomposed version of the system, we use UPPAAL to model several well-known cyber attacks and formally verify the system's behavior under these attacks. To determine viability of our approach, we first use statistical model checking SMC to assess the probabilities of success for selected attacks. Based on SMC outcomes, we use symbolic model checking to individually analyse the subsystem affected by each attack. The results obtained from this analysis are then used to demonstrate the feasibility of our approach. We demonstrate our method using an actual control system architecture provided by our industrial partner. CCS CONCEPTS • Security and privacy → Logic and verification; Denial-ofservice attacks; Security requirements.

Modeling an Edge Computing Arithmetic Framework for IoT Environments

“…It is to be mentioned that the state explosion occurs when the number of states to be modeled increases with the addition of further aspects to the model, which results in an rapid increasing number of states to be included, thus leading to a fast growing number of transitions among those states, hence making the model cumbersome [55].…”

Modeling and reliability verification of industrial control network protocol based on time state transition matrix

“…7 In addition, with the maturity of model checking techniques and tools, model checking has been widely used in the analysis and verification of various safety-critical systems. 8 Therefore, we can use model checking techniques for formal design and verification of industrial control network protocols at an early stage to discover protocol design flaws and improve the reliability of the protocol development process.…”

“…With the advent of some automated theorem provers, these problems have been alleviated to some extent, but the use of these tools still requires a high level of expertise of the personnel involved 7 . In addition, with the maturity of model checking techniques and tools, model checking has been widely used in the analysis and verification of various safety‐critical systems 8 . Therefore, we can use model checking techniques for formal design and verification of industrial control network protocols at an early stage to discover protocol design flaws and improve the reliability of the protocol development process.…”