Modified Multilevel Feedback Queue Scheduling Algorithm with Starvation Mitigation for Reconfigurable Computing systems

Abstract

Abstract Multilevel feedback queue (MLFQ) is one of the efficient scheduling algorithms used in popular operating systems such as windows and Linux. The algorithm consists of several queues with varying priorities and with the low priority queue taking care of starvation mitigation. As Reconfigurable Computing Systems(RCS) handle hardware tasks with different priorities and short and long execution times, which are to be executed in the programmable hardware, MLFQ is suitably modified to match the needs of an RCS.Unlike the traditional software implementations, the hardware tasks can not be preempted at ease and hence the algorithm is suitably modified with four queues based on priority and execution times of the tasks. The MMLFQ is implemented and its performance is compared with the other conventional scheduling algorithms such as static configuration, SJF, FCFS and MLFQ. It is found that the proposed Modified MLFQ technique provides a reduced waiting time for both short jobs of low and high priority tasks and is a better choice than other scheduling algorithms.