Trade-Off Model-Based Traffic Sampling for Intrusion Detection in Hybrid Vehicular Networks

In vehicular networks, as diverse traffic demands continue to grow, the mitigation of malfunctions in vehicular connectivity services caused by malicious data initiated by attackers has evolved into a substantial research issue. The intrusion detection system (IDS) could be implemented in vehicles to monitor malicious traffic behavior through deep packet inspections. However, the limited capacity of in-vehicle IDS makes inspecting all the incoming packets challenging. Instead, in-network IDS near the network edge could be utilized to support packet inspection during packet propagation. In this paper, we propose a traffic sampling method for packet inspection at IDS in hybrid vehicular networks, leveraging the trade-off between the probability of successful packet inspection and inspection delay. Based on the trade-off relation model, the proposed method decides the sampling rate at the software-defined networking (SDN)-enabled switches. With an appropriate adjustment of the control parameters of the proposed method, the simulation results show that the proposed method efficiently provides successful packet inspection for dynamically changing vehicular networks.