Virtual Platforms and Timing Analysis: Status, Challenges and Future Directions

This paper outlines a methodology based on virtual platforms and timing analysis to perform the exploration and selection of automotive architecture solutions. The analysis provides a quantitative evaluation of architecture options with respect to para-functional metrics. The satisfaction of deadline constraints and the optimization with respect to latency on end-to-end computations is considered. In addition, we discuss to what degree existing standards, including OSEK and AUTOSAR and model-based development practices can support the development of time predictable software and what is required to move toward the desirable goal of timing isolation when integrating multiple applications into the same execution platform. Finally, the paper provides a quick glance at recent results in the the optimization of the software architecture for complex distributed systems with respect to worst case timing behavior.