Title: Proximity-constrained and cost-efficient scheduling of workflows over edge computing resources

Authors: Yuyin Ma; Ziyang Zeng; Yunni Xia

Addresses: Software Theory and Technology Chongqing Key Lab, Chongqing University, Chongqing, China ' School of Mechanical Engineering, Chongqing University, Chongqing, China ' Software Theory and Technology Chongqing Key Lab, Chongqing University, Chongqing, China

Abstract: The edge computing paradigm is emerging as a novel paradigm that is capable of offloading computation from centralised nodes to edge resources. It provides highly cost-efficient computing resources, storage and network services near the edge. A key challenge for workflow scheduling upon edge is the reduction monetary cost while fulfilling service-level-agreement. However, it is difficult to guarantee user-perceived performance of applications deployed upon edge infrastructures because such applications are constantly subject to negative impacts, e.g., network congestions, unexpected long message delays, shrinking coverage range of edge servers due to battery depletion. In this work, we study the multi-workflow scheduling problem and develop a novel approach to cost-efficient scheduling of multi-workflows upon edge. The considered approach minimises edge computing costs while meeting user-specified workflow completion deadlines by leveraging a discrete firefly algorithm for yielding the scheduling plan. We perform experimental case studies based on multiple well-known scientific workflow templates and a real-world dataset of edge resource locations as well. Experimental results clearly suggest that our proposed approach beats traditional ones.

Keywords: edge computing; workflow scheduling; cost efficiency; proximity constraint.

DOI: 10.1504/IJIITC.2020.110282

International Journal of Intelligent Internet of Things Computing, 2020 Vol.1 No.2, pp.100 - 113

Received: 09 Sep 2019
Accepted: 28 Nov 2019
Published online: 12 Oct 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article