Abstract
In real-time systems, a task or a set of tasks needs to be executed and completed successfully within a predefined time. Those systems require a scheduling technique or a set of scheduling methods to distribute the given task or the set of tasks among different processors or on a processor. In this paper, a new novel scheduling approach to minimize the overhead from context switching between several periodic tasks is presented. This method speeds up a required response time while ensuring that all tasks meet their deadline times and there is no deadline miss occurred. It is a dynamic-priority technique that works either on a uniprocessor or several processors. In particular, it is proposed to be applied on multiprocessor environments since many applications run on several processors. Various examples are presented within this paper to demonstrate its optimality and efficiency. In addition, several comparison experiments with an earlier version of this approach were performed to demonstrate its efficiency and effectiveness too. Those experiments showed that this novel approach sped up the execution time from 15% to nearly around 46%. In addition, it proved that it reduced the number of a context switch between tasks from 12% to around 50% as shown from simulation tests. Furthermore, this approach delivered all tasks/jobs successfully and ensured there was no deadline miss happened.
Highlights
In real-time systems, a task or a set of tasks needs to be executed and completed successfully within a predefined time
Deciding which task must be selected first from several existing tasks in the ready queue and ensuring that no task misses its deadline time are considered the key purposes of any scheduling technique (Alsheikhy et al, 2016)
The CPU and resource utilization in any hard real-time systems can be affected by a scheduling method
Summary
System’s functioning which can be seen as the Quality of Service (QoS) (Guo and Baruah, 2015). Real-time systems perform several tasks with different priorities These tasks can be run on a uniprocessor environment or on multiprocessor environments (Guo and Baruah, 2015; Ren and Phan, 2015; Harkut and Agrawal, 2014). The EDF technique is a dynamic one since any task with the shortest deadline time becomes the first task in the ready queue among all other tasks This scheme can be seen as the optimal method in uniprocessor and multiple processor environments for both types of tasks “periodic and aperiodic (Alsheikhy et al, 2016).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
