On the optimal allocation of adversarial resources

Security is important for mission-critical wireless sensor networks (WSNs). This is especially so because powerful adversaries could compromise and control a significant fraction of the network nodes. A plethora of schemes has been developed to secure wireless sensor networks and resilience to sophisticated attacks has been analyzed. However, the question of how the adversary could deploy her resources to maximally affect the attacked system has remained largely unaddressed. This is the problem this paper is concerned with: Given a number of compromised entities (nodes) and cryptographic keys, how can the adversary devise a close-to-optimal attack tactic? To the best of our knowledge, this is the first investigation of its kind: while the basic adversarial behavior is well-known, the problem of how the adversary can optimally deploy her resources to maximize the attack impact has not been considered for WSNs. We consider an abstract model of the mission-critical WSN and the adversary, and we find that the determination of an optimal attack is computationally hard, thus, we devise an efficient heuristic approach. An intelligent adversarial resource allocation indeed yields disproportional gains for the attacker. Our analysis is the first necessary step to comprehend how to best address vulnerabilities.