Skip to Main Content
Soft real-time applications, such as multimedia systems, are becoming increasingly unpredictable due to the variation of the tasks' attributes. Under these circumstances, the scheduling algorithms depending on the tasks' static attributes can't provide useful and efficient scheduling support to those soft real-time systems. In this paper, we present a dynamic feedback and elastic scheduling model and period adjustment algorithm for soft real-time application with flexible workload. Based on logging the number of task instance that miss deadline periodically, this model adjusts the tasks' executing period to change the resource requirement in the next sampling period. This model makes the system state converge to the steady state when the system is overloaded, and makes tasks to use system resource sufficiently when system resource isn't fully utilized. We analyse the model and evaluate its performance.