ShuffleCAN: Enabling Moving Target Defense for Attack Mitigation on Automotive CAN

Huiping Qian,Hao Han,Xiaojun Zhu,Fengyuan Xu

DOI: https://doi.org/10.1109/msn60784.2023.00059

2023-01-01

Abstract:Controller Area Networks (CANs), the most widely used protocols for in-vehicle networks, are vulnerable to various attacks due to the lack of security countermeasures by design. CAN messages are broadcast without source/destination labeling and lack built-in encryption or authentication mechanisms, thus suffering many attacks. To address this problem, we propose a lightweight CAN message obfuscation technique called ShuffleCAN. Motivated by the idea of moving target defense (MTD), ShuffleCAN is designed with a combined shuffling scheme based on the hash chain and combinatorial coding techniques to achieve both ID anonymization and payload shuffling. With ShuffleCAN, selected or all transmitter and receiver pairs can communicate in a private dialect over the standard CAN protocol, so the eavesdropper cannot understand the meaning of each message or inject a valid fake message. We implemented a prototype and evaluated ShuffleCAN on Toyota’s testbed PASTA. The experimental results show that ShuffleCAN outperforms state-of-the-art CAN protection schemes.