RESOURCE RECLAIMING IN MULTIPROCESSOR REAL-TIME SYSTEMS

Most real-time scheduling algorithms schedule tasks with respect to th eir worst case computation times. Resource reclaiming refers to the pr oblem of utilizing the resources left unused by a task when it execute s less than its worst case computation time, or when a task is deleted from the current schedule. Resource reclaiming is a very important is sue in dynamic real-time multiprocessor environments. In this paper, w e present dynamic resource reclaiming algorithms that are effective, a void any run time anomalies, and have bounded overhead cost that is in dependent of the number of tasks in the schedule. Each Task is assumed to have a worst case computation time, a deadline, and a set of resou rce requirements. The algorithms utilize the information given in a mu ltiprocessor task schedule and perform on-line local optimization. The effectiveness of the algorithms is demonstrated through simulation st udies. The algorithms have also been implemented in the Spring Kernel [15].