Results on Competitiveness of Online Shortest Remaining Processing Time(SRPT) Scheduling with Special Classes of Inputs

TL;DR

This paper investigates the performance of the SRPT scheduling algorithm in online multiprocessor systems focusing on minimizing makespan, providing improved competitiveness results for specific real-world job sequences.

Contribution

It offers new competitiveness bounds for SRPT in online multiprocessor scheduling with makespan minimization for certain input classes.

Findings

Improved constant competitiveness results for SRPT.

Analysis specific to special classes of online job sequences.

Application relevance to real-life scheduling scenarios.

Abstract

Shortest Remaining Processing Time (SRPT) is a well known preemptive scheduling algorithm for uniprocessor and multiprocessor systems. SRPT finds applications in the emerging areas such as scheduling of client's requests that are submitted to a web server for accessing static web pages, managing the access requests to files in multiuser database systems and routing of packets across several links as per bandwidth availability in data communications. SRPT has been proved to be optimal for the settings, where the objective is to minimize the mean response time of a list of jobs. According to our knowledge, there is less attention on the study of online SRPT with respect to the minimization of makespan as a performance criterion. In this paper, we study the SRPT algorithm for online scheduling in multiprocessor systems with makespan minimization as an objective. We obtain improved constant competitiveness results for algorithm SRPT for special classes of online job sequences based on practical real life applications.