Information Fusion to Defend Intentional Attack in Internet of Things

Abstract

Robust network design against attacks is one of the most fundamental issues in Internet of Things (IoT) architecture as IoT operations highly rely on the support of the underlaying communication infrastructures. In this paper, the vulnerability of IoT infrastructure under intentional attacks is investigated by relating the network resilience to the percolation-based connectivity. Intentional attacks impose severe threats on the network operations as it can effectively disrupt a network by paralyzing a small fraction of nodes, and therefore deteriorating IoT operations. A fusion-based defense mechanism is proposed to mitigate the damage caused by such attacks, where each node feedbacks minimum (one-bit) local decision to the fusion center for attack inference. By formulating the attack and defense strategy as a zero-sum game, the outcome of the game equilibrium is used to evaluate the effectiveness of the proposed mechanism. The robustness of the Internet-oriented and the cyber-physical system (CPS)-oriented networks are specifically analyzed to illustrate the foundation of future IoT infrastructure. Both analytical and empirical results show that the proposed mechanism greatly enhances the robustness of IoT, even in the weak local detection capability and fragile network structure regime.