CEVP: Cross Entropy based Virtual Machine Placement for Energy Optimization in Clouds

Xiaodao Chen; Yunliang Chen; Albert Y. Zomaya; Rajiv Ranjan; Shiyan Hu

doi:10.1007/s11227-016-1630-1

CEVP: Cross Entropy based Virtual Machine Placement for Energy Optimization in Clouds

Xiaodao Chen, Yunliang Chen^*, Albert Y. Zomaya, Rajiv Ranjan, Shiyan Hu

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

21 Citations (Scopus)

Abstract

Big data trends have recently brought unrivalled opportunities to the cloud systems. Numerous virtual machines (VMs) have been widely deployed to enable the on-demand provisioning and pay-as-you-go services for customers. Due to the large complexity of the current cloud systems, promising VM placement algorithm are highly desirable. This paper focuses on the energy efficiency and thermal stability issues of the cloud systems. A Cross Entropy based VM Placement (CEVP) algorithm is proposed to simultaneously minimize the energy cost, total thermal cost and the number of hot spots in the data center. Simulation results indicate that the proposed CEVP algorithm can (1) achieve energy savings of 26.2 % on average, (2) efficiently reduce the temperature cost by up to 6.8 % and (3) significantly decrease the total number of the hot spots by 60.1 % on average in the cloud systems, by comparing to the Ant Colony System-based algorithm.

Original language	English
Pages (from-to)	3194-3209
Number of pages	16
Journal	Journal of Supercomputing
Volume	72
Issue number	8
DOIs	https://doi.org/10.1007/s11227-016-1630-1
Publication status	Published - 1 Aug 2016
Externally published	Yes

Access to Document

10.1007/s11227-016-1630-1

Cite this

@article{8417d6fcf5f147d2945d3e060bfad6a3,

title = "CEVP: Cross Entropy based Virtual Machine Placement for Energy Optimization in Clouds",

abstract = "Big data trends have recently brought unrivalled opportunities to the cloud systems. Numerous virtual machines (VMs) have been widely deployed to enable the on-demand provisioning and pay-as-you-go services for customers. Due to the large complexity of the current cloud systems, promising VM placement algorithm are highly desirable. This paper focuses on the energy efficiency and thermal stability issues of the cloud systems. A Cross Entropy based VM Placement (CEVP) algorithm is proposed to simultaneously minimize the energy cost, total thermal cost and the number of hot spots in the data center. Simulation results indicate that the proposed CEVP algorithm can (1) achieve energy savings of 26.2 % on average, (2) efficiently reduce the temperature cost by up to 6.8 % and (3) significantly decrease the total number of the hot spots by 60.1 % on average in the cloud systems, by comparing to the Ant Colony System-based algorithm.",

keywords = "Cloud computing, Optimization, VM placement",

author = "Xiaodao Chen and Yunliang Chen and Zomaya, {Albert Y.} and Rajiv Ranjan and Shiyan Hu",

note = "Publisher Copyright: {\textcopyright} 2016, Springer Science+Business Media New York.",

year = "2016",

month = aug,

day = "1",

doi = "10.1007/s11227-016-1630-1",

language = "English",

volume = "72",

pages = "3194--3209",

journal = "Journal of Supercomputing",

issn = "0920-8542",

publisher = "Springer Netherlands",

number = "8",

}

TY - JOUR

T1 - CEVP

T2 - Cross Entropy based Virtual Machine Placement for Energy Optimization in Clouds

AU - Chen, Xiaodao

AU - Chen, Yunliang

AU - Zomaya, Albert Y.

AU - Ranjan, Rajiv

AU - Hu, Shiyan

PY - 2016/8/1

Y1 - 2016/8/1

N2 - Big data trends have recently brought unrivalled opportunities to the cloud systems. Numerous virtual machines (VMs) have been widely deployed to enable the on-demand provisioning and pay-as-you-go services for customers. Due to the large complexity of the current cloud systems, promising VM placement algorithm are highly desirable. This paper focuses on the energy efficiency and thermal stability issues of the cloud systems. A Cross Entropy based VM Placement (CEVP) algorithm is proposed to simultaneously minimize the energy cost, total thermal cost and the number of hot spots in the data center. Simulation results indicate that the proposed CEVP algorithm can (1) achieve energy savings of 26.2 % on average, (2) efficiently reduce the temperature cost by up to 6.8 % and (3) significantly decrease the total number of the hot spots by 60.1 % on average in the cloud systems, by comparing to the Ant Colony System-based algorithm.

AB - Big data trends have recently brought unrivalled opportunities to the cloud systems. Numerous virtual machines (VMs) have been widely deployed to enable the on-demand provisioning and pay-as-you-go services for customers. Due to the large complexity of the current cloud systems, promising VM placement algorithm are highly desirable. This paper focuses on the energy efficiency and thermal stability issues of the cloud systems. A Cross Entropy based VM Placement (CEVP) algorithm is proposed to simultaneously minimize the energy cost, total thermal cost and the number of hot spots in the data center. Simulation results indicate that the proposed CEVP algorithm can (1) achieve energy savings of 26.2 % on average, (2) efficiently reduce the temperature cost by up to 6.8 % and (3) significantly decrease the total number of the hot spots by 60.1 % on average in the cloud systems, by comparing to the Ant Colony System-based algorithm.

KW - Cloud computing

KW - Optimization

KW - VM placement

UR - http://www.scopus.com/inward/record.url?scp=84957535003&partnerID=8YFLogxK

U2 - 10.1007/s11227-016-1630-1

DO - 10.1007/s11227-016-1630-1

M3 - Article

SN - 0920-8542

VL - 72

SP - 3194

EP - 3209

JO - Journal of Supercomputing

JF - Journal of Supercomputing

IS - 8

ER -

CEVP: Cross Entropy based Virtual Machine Placement for Energy Optimization in Clouds

Abstract

Access to Document

Other files and links

Fingerprint

Cite this