Virtualizing an Automotive State-of-the-Art Microcontroller: Techniques and Its Evaluation

The future of automotive industry has a major focus on computing innovations such as autonomous driving, connectivity, and mobility. With these advancements, electrical and electronic components started exponentially increasing inside the vehicle, integrating hardware and software components of different automotive safety integrity levels. Automotive OEMs and their suppliers are seeking for innovative and optimized electrical/electronic vehicle architecture to improve vehicle performance, safety, reliability, and lower system costs. Consolidating several small heterogeneous computing units to a centralized computing unit is an approach for optimizing the electrical/electronic vehicle architecture. A demonstrator virtualizing an automotive state-of-the-art multicore controller with two heterogeneous hard real-time applications is realized. Through this demonstrator, various new concepts like start-up of virtualized system, trap-emulation, virtualizing input–output access, and interrupt handling are realized. These concepts are validated in terms of performance, data consistency, memory consumption, timing to its deadlines, and reliability of the system. Compared to the research works done so far, this evaluation is based on a demonstrator where both virtualized applications are performing their regular system activities. There is no master-slave concept in this demonstrator, enabling independent access for each application to its needed peripheral.