Dynamic resource provisioning for workflow scheduling under uncertainty in edge computing environment

Summary Edge computing, an extension of cloud computing, is introduced to provide sufficient computing and storage resources for mobile devices. Moreover, a series of computing tasks in a mobile device are set as structured computing processes and flows to achieve effective management by the workflo...

Full description

Saved in:
Bibliographic Details
Published in:Concurrency and computation 2022-06, Vol.34 (14), p.n/a
Main Authors: Xu, Xiaolong, Cao, Hao, Geng, Qingfan, Liu, Xihua, Dai, Fei, Wang, Chuanjian
Format: Article
Language:English
Subjects:
Cloud computing
Completion time
Edge computing
Electronic devices
Energy consumption
Genetic algorithms
Performance degradation
Provisioning
Resource allocation
Resource scheduling
SDN
Software-defined networking
Sorting algorithms
Uncertainty
Workflow
workflow scheduling
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Summary Edge computing, an extension of cloud computing, is introduced to provide sufficient computing and storage resources for mobile devices. Moreover, a series of computing tasks in a mobile device are set as structured computing processes and flows to achieve effective management by the workflow. However, the execution uncertainty caused by performance degradation, service failure, and new service additions remains a huge challenge to the user's service experience. In order to address the uncertainty, a software‐defined network (SDN)‐based edge computing framework and a dynamic resource provisioning (UARP) method are proposed in this paper. The UARP method is implemented in the proposed framework and addresses the uncertainty through the advantages of SDN. In addition, the nondominated sorting genetic algorithm‐III is employed to optimize two goals, that is, the energy consumption and the completion time, to obtain balanced scheduling strategies. The comparative experiments are performed and the results show that the UARP method is superior to other methods in addressing the uncertainty, while reducing energy consumption and shortening the completion time.
ISSN:1532-0626
1532-0634
DOI:10.1002/cpe.5674