Efficient Post-Silicon Run-Time Error Detection for Systems-on-Chip (Extended Abstract)

Embedded systems have become very important components in many safety-critical areas. Because of that, methods that ensure a correct design and execution of those systems are crucial. How to guarantee that also after deployment remains a challenging task. Current state-of-the-art solutions try to address this problem by adding redundant hardware and comparing the respectively obtained results. However, although such methods are commonly used in industry, they lead to embedded systems with huge additional costs with respect to size, power consumption, and delay. In this work, we are proposing a methodology for error detection after deployment that works without this overhead. To this end, we utilize techniques for program analysis which allows for extracting information that can be used by the embedded system to check whether it is still in a correct state. Case studies conducted within an industrial environment confirmed the benefits of the proposed methodology.