Network security model for analyzing network-based control systems under denial of service attacks

Abstract

Denial of service (DoS) attacks have become major threats to network security, which is pertinent to the deployment and performance of network-based control systems (NBCS). In this paper, we propose two queueing models to simulate the stochastic process of packet transmission under DoS attacks. The motivation is to quantitatively investigate how the attacks affect the performance of NBCS. The control system consists of a discrete PI controller (either event- or time-driven), a second-order plant, and two one-way delay vectors induced by networks. Experimental results indicate that the event-driven controller is more robust than the time-driven one under attacks. Model I DoS attacks (excessive packet loss) impair the performance, but do not destabilize the system with the event-driven controller. Model II DoS attacks (increased delay jitter) deteriorate the performance or even destabilize the system.