To meet its timing requirements, a real-time system must adopt a scheduling algorithm that permits to ensure a predictable response time. In this paper, we are specifically concerned with the problem of serving soft sporadic tasks in a uniprocessor system in which hard periodic tasks are scheduled using a dynamic priority algorithm. Typically, soft sporadic tasks benefit from being executed as early as possile while periodic tasks need to meet their deadlines. First, we present an analysis of preemptive Earliest Deadline scheduling for a model of independent periodic tasks. Then, we augment our analysis to cater for periodic tasks which exclusively access to critical sections handled by the Kemelized Monitor protocol. Our analysis determines the maximum processing lime which may be stolen from periodic tasks without jeopardizing both their timing constraints and resource consistency. It provides the basis for an optimal dynamic scheduling algorithm in a context of resource sharing. Optimality means that every soft sporadic task is executed with a minimal response time.
Real-time systems, Dynamic scheduling, Hard deadline periodic tasks, Sporadic tasks, Resource constraints.
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).