An Adaptive Mechanism for Reducing Criticality Level in Mixed-Criticality Systems

How to raise criticality level to avoid miss the deadlines of high critical tasks has been discussed a lot, then how to reduce criticality level to assure multiple functional tasks efficiently using resource on a platform is less researched. In this paper, we present an adaptive mechanism, which is based on evaluating high-critical tasks execution time and feedback to admission controller with dynamic priority scheduling. Comparing the evaluated utilization with the low critical characteristic of high critical task, it determine the exact time point to reduce criticality level safely and effectively, meeting the deadlines of high critical tasks as well as correctly running low critical tasks.