A improved elastic scheduling algorithm based on feedback control theory

Due to the variable of the tasks´ attributes, the behavior of soft real-time systems is becoming increasingly unpredictable. Under this circumstance, the scheduling algorithms, which depend on the tasks´ static attributes, can´t provide usable and efficient resource allocation for those soft real-time systems. In this paper, we present an elastic scheduling algorithm for flexible workload firstly. Based on logging system resource utilization and the number of task instances executing and lost in a sampling period, this algorithm adjusts the number of task instances executing in the next sampling period to guarantee the tasks´ basic QoS and improve the system resource utilization. Based on the above algorithm, we establish feedback control model for controlled system and controller, and use feedback control algorithm to improve the efficiency and stability for resource allocation. In this paper, we analyze the algorithms, and simulate them on NS2, and evaluate their performance.