Visible to the public Security games on infrastructure networks

TitleSecurity games on infrastructure networks
Publication TypeConference Paper
Year of Publication2016
AuthorsAmin, S.
Conference Name2016 Science of Security for Cyber-Physical Systems Workshop (SOSCYPS)
Date Publishedapr
ISBN Number978-1-5090-4304-0
Keywordsattacker-defender interactions, Collaboration, communication failures, Communication networks, composability, controller-disturbance interaction, Cyber-physical security, Cyber-physical systems, cyber-security modelling, Density estimation robust algorithm, distribution networks, game theory, game theory methods, Games, Human Behavior, human factors, infrastructure networks, Inspection, interdependent security game, interdiction problem, Investment, IT systems, Load modeling, Metrics, network security game, networked control systems, node disruptions, optimal attack plans, Policy-Governed Secure Collaboration, pubcrawl, Resiliency, resource allocation, robust control models, Scalability, science of security, security, security of data, strategic edge disruptions

The theory of robust control models the controller-disturbance interaction as a game where disturbance is nonstrategic. The proviso of a deliberately malicious (strategic) attacker should be considered to increase the robustness of infrastructure systems. This has become especially important since many IT systems supporting critical functionalities are vulnerable to exploits by attackers. While the usefulness of game theory methods for modeling cyber-security is well established in the literature, new game theoretic models of cyber-physical security are needed for deriving useful insights on "optimal" attack plans and defender responses, both in terms of allocation of resources and operational strategies of these players. This whitepaper presents some progress and challenges in using game-theoretic models for security of infrastructure networks. Main insights from the following models are presented: (i) Network security game on flow networks under strategic edge disruptions; (ii) Interdiction problem on distribution networks under node disruptions; (iii) Inspection game to monitor commercial non-technical losses (e.g. energy diversion); and (iv) Interdependent security game of networked control systems under communication failures. These models can be used to analyze the attacker-defender interactions in a class of cyber-physical security scenarios.

Citation Keyamin_security_2016