A Novel Federated Earliest Deadline FirstReal-time Task Scheduler forMultiprocessors

This research paper focuses on the development of scheduling algorithms for real-time systems, with aspecific emphasis on multiprocessor environments. The primary objective is to ensure the timely execution ofhard real-time tasks to avoid any detrimental consequences caused by missed deadlines. To achieve thisgoal, Federated Earliest Deadline First (fEDF) scheduler is proposed, which exhibits promising capabilitiesin meeting the desired objective. The fEDF algorithm incorporates shared and dedicated modes, enhancingflexibility and minimizing idle time and overhead through the utilization of shared tasks. Experimental resultsvalidate the effectiveness of the proposed algorithm, demonstrating its ability to significantly improveutilization and maximize processor efficiency while maintaining comparable overhead to the singleprocessormode. This research contributes to the advancement of scheduling algorithms for multiprocessorsystems, effectively addressing the challenges associated with reliable and efficient real-time task execution

Keywords:

Federated EDF , Semi-partitioned scheduler , Hard real-time scheduler , Real-time periodictasks , Multiprocessors