Event-based utilization control for distributed Teal-time Embedded systems

Distributed Real-time Embedded (DRE) systems working in open environments often need to enforce utilization bounds on multiple processors in order to manage CPU resources adaptively even when task execution times are unforeseeable. An event-based utilization control algorithm is proposed, and the algorithm applies to mixed (continuous and discrete) rate adaptation in DRE systems including tasks that support discrete configurations and the ones that support continuous configurations. A mixed integer predictive controller is designed and it is triggered when needed so that overhead can be reduced greatly. The experimental results argue the effectiveness of this algorithm.