Computing Sciences and Computer Engineering
Nowadays, sensors, communications connections, and more powerful computing capabilities are added to automobiles, making them more intelligent. The primary goal was to eliminate the need for human control, making them Autonomous Vehicles (AVs). Consequently, researchers thought to put all that newly added computational power to use for other endeavors. Hence, Autonomous Vehicular Cloud Computing (AVCC) models were introduced. Nevertheless, this goal is not an easy undertaking, the dynamic nature of autonomous vehicles introduces a critical challenge in the development of such a distributed computing platform. Furthermore, it presents far complicated issues as far as security and protection of services associated with this framework. In this paper, we center around securing programs running on AVCC. Here, we focus on timing side-channel attacks which aim to leak information about running code, which can be utilized to reverse engineer the program itself. We propose to mitigate these attacks via obfuscated compilation. In particular, we change the control flow of an input program at the compiler level, thereby changing the program’s apparent behavior and accompanying physical manifestations to hinder these attacks. We improve our previous ARM-based implementation to address its limitations and provide more comprehensive coverage for different programs. Our solution is software-based and generically portable - fitting different hardware platforms and numerous input program languages at the source level. Our findings prove a considerable improvement over our previous technique, which may provide more defense against timing side-channels.
(2022). Enhanced Obfuscation for Software Protection in Autonomous Vehicular Cloud Computing Platforms. IEEE Access, 10, 33943-33953.
Available at: https://aquila.usm.edu/fac_pubs/19955