Two approaches towards EDZL scheduling for performance asymmetric multiprocessors

Abstract

In order to improve the performance of multi-threaded applications for real-time systems such as network servers and multimedia systems, asymmetric multiprocessors have been proposed. The benefits of improved performance and reduced power consumption from such architectures cannot be fully exploited unless suitable scheduling and task allocation methods are implemented at the operating system level. Our current research focuses on providing efficient scheduling algorithm for performance asymmetric multiprocessors used in real-time applications. Specifically, we present two approaches for real-time task allocation based on EDZL scheduling policy depending on the choice of speed of processors. The first approach chooses a fastest speed processor for high priority tasks. The second approach chooses a slowest speed processor for higher priority non-zero laxity tasks. We explain these two scheduling methods with examples and also derive schedulability tests for both approaches.