Jamming-based adversarial control of network flow allocation: A passivity approach

Abstract

Wireless cyber-physical systems are vulnerable to jamming attacks, in which an adversary broadcasts an interfering signal in the vicinity of a receiver, causing packet decoding errors and reducing the throughput of the communication. Reduced throughput and increased delay could violate the realtime constraints of cyber-physical systems. In a flow redirection attack, an adversary jams a set of network links in order to cause network sources to divert traffic to links that are controlled by the adversary, enabling higher-layer attacks. In this paper, we introduce a passivity approach for modeling the flow redirection attack. Using our approach, we identify a class of dynamic jamming strategies for flow redirection, in which the adversary updates the probability of jamming based on the rate of flow traversing the link. We provide sufficient conditions for feasibility of the jamming strategies for energy-constrained adversaries, and develop an efficient algorithm for deriving an optimal jamming strategy for a given network and desired flow allocation. Our results are illustrated via a numerical study.