Multi-objective auto-scaling scheduling for micro-service workflows in hybrid clouds

Shijia Wang; Xuan Liu; Ming Gao; Mingxia Chen; Kai Leung Yung; Shancheng Jiang

doi:10.1080/17517575.2022.2069478

Multi-objective auto-scaling scheduling for micro-service workflows in hybrid clouds

Shijia Wang, Xuan Liu, Ming Gao, Mingxia Chen, Kai Leung Yung, Shancheng Jiang

Research output: Journal article publication › Journal article › Academic research › peer-review

2 Citations (Scopus)

Abstract

A novel multi-objective (cost, delay, and reliability) autoscaling optimisation model is proposed for micro-service workflows in containerised hybrid clouds. We compare the container based model with VM-based model and conclude that the former significantly supersedes. The benchmark of three mainstream algorithms is conducted by the Hypervolume metric, showed that the performance of MOEA/D is inferior toNSGA family, and NSGA-III is not always superior to NSGA-II. So we design an improved NSGAII based on dynamically changing crossover and mutation operators, which outperforms NSGA-III both in stability and performance by over 60% and 80% in all multi-scale tests.

Original language	English
Article number	2069478
Pages (from-to)	941-980
Number of pages	40
Journal	Enterprise Information Systems
Volume	17
Issue number	7
DOIs	https://doi.org/10.1080/17517575.2022.2069478
Publication status	Published - 3 Jul 2023

Keywords

Hybrid clouds
NSGA-II
NSGA-III
micro-service workflow scheduling
multi-objective optimisation

ASJC Scopus subject areas

Computer Science Applications
Information Systems and Management

More information

10.1080/17517575.2022.2069478

Cite this

@article{6678b753d2f8440790832a00d159206a,

title = "Multi-objective auto-scaling scheduling for micro-service workflows in hybrid clouds",

abstract = "A novel multi-objective (cost, delay, and reliability) autoscaling optimisation model is proposed for micro-service workflows in containerised hybrid clouds. We compare the container based model with VM-based model and conclude that the former significantly supersedes. The benchmark of three mainstream algorithms is conducted by the Hypervolume metric, showed that the performance of MOEA/D is inferior toNSGA family, and NSGA-III is not always superior to NSGA-II. So we design an improved NSGAII based on dynamically changing crossover and mutation operators, which outperforms NSGA-III both in stability and performance by over 60% and 80% in all multi-scale tests.",

keywords = "Hybrid clouds, NSGA-II, NSGA-III, micro-service workflow scheduling, multi-objective optimisation",

author = "Shijia Wang and Xuan Liu and Ming Gao and Mingxia Chen and Yung, {Kai Leung} and Shancheng Jiang",

note = "Publisher Copyright: {\textcopyright} 2022 Informa UK Limited, trading as Taylor & Francis Group.",

year = "2023",

month = jul,

day = "3",

doi = "10.1080/17517575.2022.2069478",

language = "English",

volume = "17",

pages = "941--980",

journal = "Enterprise Information Systems",

issn = "1751-7575",

publisher = "Taylor and Francis Ltd.",

number = "7",

}

TY - JOUR

T1 - Multi-objective auto-scaling scheduling for micro-service workflows in hybrid clouds

AU - Wang, Shijia

AU - Liu, Xuan

AU - Gao, Ming

AU - Chen, Mingxia

AU - Yung, Kai Leung

AU - Jiang, Shancheng

PY - 2023/7/3

Y1 - 2023/7/3

N2 - A novel multi-objective (cost, delay, and reliability) autoscaling optimisation model is proposed for micro-service workflows in containerised hybrid clouds. We compare the container based model with VM-based model and conclude that the former significantly supersedes. The benchmark of three mainstream algorithms is conducted by the Hypervolume metric, showed that the performance of MOEA/D is inferior toNSGA family, and NSGA-III is not always superior to NSGA-II. So we design an improved NSGAII based on dynamically changing crossover and mutation operators, which outperforms NSGA-III both in stability and performance by over 60% and 80% in all multi-scale tests.

AB - A novel multi-objective (cost, delay, and reliability) autoscaling optimisation model is proposed for micro-service workflows in containerised hybrid clouds. We compare the container based model with VM-based model and conclude that the former significantly supersedes. The benchmark of three mainstream algorithms is conducted by the Hypervolume metric, showed that the performance of MOEA/D is inferior toNSGA family, and NSGA-III is not always superior to NSGA-II. So we design an improved NSGAII based on dynamically changing crossover and mutation operators, which outperforms NSGA-III both in stability and performance by over 60% and 80% in all multi-scale tests.

KW - Hybrid clouds

KW - NSGA-II

KW - NSGA-III

KW - micro-service workflow scheduling

KW - multi-objective optimisation

UR - http://www.scopus.com/inward/record.url?scp=85132667917&partnerID=8YFLogxK

U2 - 10.1080/17517575.2022.2069478

DO - 10.1080/17517575.2022.2069478

M3 - Journal article

SN - 1751-7575

VL - 17

SP - 941

EP - 980

JO - Enterprise Information Systems

JF - Enterprise Information Systems

IS - 7

M1 - 2069478

ER -

Multi-objective auto-scaling scheduling for micro-service workflows in hybrid clouds

Abstract

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this