Improving MCT scheduling algorithm to reduce the makespan and cost of workflow execution in the cloud

Lookup NU author(s): Faris Llwaah, Dr Nigel Thomas, Dr Jacek Cala ORCiD

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.

Abstract

Cloud computing has become a modern paradigm that delivers IT resources as a service over the internet on the pay-per-use basis. Although scientific workflow applications can benefit from resources provided by cloud computing, the challenge is to find scheduling algorithms that can optimise workflow execution. This paper proposes MCT+, an extension to the Minimum Completion Time (MCT) algorithm that can reduce the makespan and cost of workflow execution. We evaluate our algorithm using WorkflowSim and two workflows from the Pegasus workflow management system. Our results show that MCT+ can improve MCT by about 35%.

Publication metadata

Author(s): Llwaah F, Thomas N, Cala J

Editor(s): Djemame, K;Kavanagh, R;Armstrong, D;

Publication type: Conference Proceedings (inc. Abstract)

Publication status: Published

Conference Name: UK Performance Engineering Workshop

Year of Conference: 2015

Pages: 1-8

Acceptance date: 17/07/2015

Publisher: University of Leeds

ePrints