Abstract
Task scheduling has a significant impact on the performance of the MapReduce computing framework. In this paper, a scheduling problem of periodical batch jobs with makespan minimization is considered. The problem is modeled as a general two-stage hybrid flow shop scheduling problem with schedule-dependent setup times. The new model incorporates the data locality of tasks and is formulated as an integer program. Three heuristics are developed to solve the problem and an improvement policy based on data locality is presented to enhance the methods. A lower bound of the makespan is derived. 150 instances are randomly generated from data distributions drawn from a real cluster. The parameters involved in the methods are set according to different cluster setups. The proposed heuristics are compared over different numbers of jobs and cluster setups. Computational results show that the performance of the methods is highly dependent on both the number of jobs and the cluster setups. The proposed improvement policy is effective and the impact of the input data distribution on the policy is analyzed and tested.
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 callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
