An optimal algorithm for guaranteeing sporadic tasks in hard real-time systems

Abstract

Guaranteeing that time critical tasks will meet their timing constraints is an important aspect of real-time systems research. Indeed, a real time system is dynamic and consequently requires online and adaptive scheduling strategies. The authors survey and extend results on scheduling hard deadline periodic and sporadic tasks on the monoprocessor model. The periodic tasks are independent, run cyclically and their characteristics are known in advance. The authors allow for the unpredictable arrival of aperiodic tasks said to be sporadic with hard real-time constraints. The problem is to jointly schedule the periodic and sporadic tasks so that the timing requirements for both sets of tasks are met. A schedulability test is presented that decides whether a new occurring task can be accepted. This test is built upon the Earliest Deadline scheduling algorithm which features a high performance and ease of implementation. >