Using ID-Hopping to Defend Against Targeted DoS on CAN

With the exponential growth of automotive security research, new security vulnerabilities and attacks have been revealed and new challenges have emerged. In recent years, various attacks ranging from replay attacks, through false information injection, to Denial of Service (DoS), have shown how fragile automotive security is. As a result, a number of security solutions have been proposed that rely on techniques like encryption and firewalls. However, most proposals require performance and computational overheads that would become an additional burden rather than a solution. In this paper, we propose a new automotive network algorithm, called ID-Hopping, that aims to prevent targeted DoS attacks in which attackers target certain functions by injecting special frames that would prevent a car's normal operations. We aim to raise the bar for attackers by randomizing the expected patterns in the automotive network. Such randomization hinders the attacker's ability to launch targeted DoS attacks. We built a testing platform and implemented the randomization mechanism to evaluate the algorithm's effectiveness. Based on the evaluation, the algorithm holds a promising solution for targeted DoS, and even reverse engineering, which automotive networks are most vulnerable to.