MMP: A Dynamic Routing Protocol Design to Proactively Defend against Wireless Network Inference Attacks
Jinmiao Chen, Zhengping Luo, Yuchen Liu, Shangqing Zhao · 2023
Network inference refers to the process of extracting sensitive information from a network without directly accessing it. This poses a significant threat to network security since it allows attackers to gain insight into sensitive information such as flow information through inference. Possessing flow information about a wireless network can empower attackers to launch more sophisticated and targeted attacks. Network inference relies on consistent traffic patterns or behavior to establish the relationship between the measured link metrics and flow information. Therefore, dynamic routing can help enhance resilience against network inference by proactive introducing variability into network traffic patterns, which can incur a high probability of mismatch between the observed patterns and the actual ones. In this paper, we observe that the inference error is positively related to the mismatch. Therefore, we propose a dynamic routing protocol, called Max-Mismatch-Probability (MMP), which seeks to maximize mismatch probability and increase the inference error. In this paper, we provide the theoretical analysis of our proposed protocol and show that the inference error of MMP is Θ(√N), which is verified in our experimental results.