Dynamic energy-aware capacity provisioning for cloud computing environments

Qi Zhang, Mohamed Faten Zhani, Shuo Zhang, Quanyan Zhu, Raouf Boutaba, Joseph L. Hellerstein

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Data centers have recently gained significant popularity as a cost-effective platform for hosting large-scale service appli- cations. While large data centers enjoy economies of scale by amortizing initial capital investment over large number of machines, they also incur tremendous energy cost in terms of power distribution and cooling. An effective approach for saving energy in data centers is to adjust dynamically the data center capacity by turning off unused machines. How- ever, this dynamic capacity provisioning problem is known to be challenging as it requires a careful understanding of the resource demand characteristics as well as considerations to various cost factors, including task scheduling delay, ma- chine reconfiguration cost and electricity price fluctuation. In this paper, we provide a control-theoretic solution to the dynamic capacity provisioning problem that minimizes the total energy cost while meeting the performance objec- tive in terms of task scheduling delay. Specifically, we model this problem as a constrained discrete-time optimal control problem, and use Model Predictive Control (MPC) to find the optimal control policy. Through extensive analysis and simulation using real workload traces from Google's compute clusters, we show that our proposed framework can achieve significant reduction in energy cost, while maintaining an acceptable average scheduling delay for individual tasks.

Original languageEnglish (US)
Title of host publicationICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing
Pages145-154
Number of pages10
DOIs
StatePublished - 2012
Event9th ACM International Conference on Autonomic Computing, ICAC'12 - San Jose, CA, United States
Duration: Sep 18 2012Sep 20 2012

Other

Other9th ACM International Conference on Autonomic Computing, ICAC'12
CountryUnited States
CitySan Jose, CA
Period9/18/129/20/12

Fingerprint

Cloud computing
Cloud Computing
Data Center
Costs
Energy
Task Scheduling
Scheduling
Power Distribution
Large Data
Model predictive control
Model Predictive Control
Control Policy
Energy Saving
Optimal Policy
Reconfiguration
Electricity
Workload
Cooling
Optimal Control Problem
Energy conservation

Keywords

  • Cloud computing
  • Energy manage- ment
  • Model predictive control
  • Resource management

ASJC Scopus subject areas

  • Artificial Intelligence
  • Applied Mathematics

Cite this

Zhang, Q., Zhani, M. F., Zhang, S., Zhu, Q., Boutaba, R., & Hellerstein, J. L. (2012). Dynamic energy-aware capacity provisioning for cloud computing environments. In ICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing (pp. 145-154) https://doi.org/10.1145/2371536.2371562

Dynamic energy-aware capacity provisioning for cloud computing environments. / Zhang, Qi; Zhani, Mohamed Faten; Zhang, Shuo; Zhu, Quanyan; Boutaba, Raouf; Hellerstein, Joseph L.

ICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing. 2012. p. 145-154.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zhang, Q, Zhani, MF, Zhang, S, Zhu, Q, Boutaba, R & Hellerstein, JL 2012, Dynamic energy-aware capacity provisioning for cloud computing environments. in ICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing. pp. 145-154, 9th ACM International Conference on Autonomic Computing, ICAC'12, San Jose, CA, United States, 9/18/12. https://doi.org/10.1145/2371536.2371562
Zhang Q, Zhani MF, Zhang S, Zhu Q, Boutaba R, Hellerstein JL. Dynamic energy-aware capacity provisioning for cloud computing environments. In ICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing. 2012. p. 145-154 https://doi.org/10.1145/2371536.2371562
Zhang, Qi ; Zhani, Mohamed Faten ; Zhang, Shuo ; Zhu, Quanyan ; Boutaba, Raouf ; Hellerstein, Joseph L. / Dynamic energy-aware capacity provisioning for cloud computing environments. ICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing. 2012. pp. 145-154
@inproceedings{479dfb7313c64935af3b9d5f0e81fc82,
title = "Dynamic energy-aware capacity provisioning for cloud computing environments",
abstract = "Data centers have recently gained significant popularity as a cost-effective platform for hosting large-scale service appli- cations. While large data centers enjoy economies of scale by amortizing initial capital investment over large number of machines, they also incur tremendous energy cost in terms of power distribution and cooling. An effective approach for saving energy in data centers is to adjust dynamically the data center capacity by turning off unused machines. How- ever, this dynamic capacity provisioning problem is known to be challenging as it requires a careful understanding of the resource demand characteristics as well as considerations to various cost factors, including task scheduling delay, ma- chine reconfiguration cost and electricity price fluctuation. In this paper, we provide a control-theoretic solution to the dynamic capacity provisioning problem that minimizes the total energy cost while meeting the performance objec- tive in terms of task scheduling delay. Specifically, we model this problem as a constrained discrete-time optimal control problem, and use Model Predictive Control (MPC) to find the optimal control policy. Through extensive analysis and simulation using real workload traces from Google's compute clusters, we show that our proposed framework can achieve significant reduction in energy cost, while maintaining an acceptable average scheduling delay for individual tasks.",
keywords = "Cloud computing, Energy manage- ment, Model predictive control, Resource management",
author = "Qi Zhang and Zhani, {Mohamed Faten} and Shuo Zhang and Quanyan Zhu and Raouf Boutaba and Hellerstein, {Joseph L.}",
year = "2012",
doi = "10.1145/2371536.2371562",
language = "English (US)",
isbn = "9781450315203",
pages = "145--154",
booktitle = "ICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing",

}

TY - GEN

T1 - Dynamic energy-aware capacity provisioning for cloud computing environments

AU - Zhang, Qi

AU - Zhani, Mohamed Faten

AU - Zhang, Shuo

AU - Zhu, Quanyan

AU - Boutaba, Raouf

AU - Hellerstein, Joseph L.

PY - 2012

Y1 - 2012

N2 - Data centers have recently gained significant popularity as a cost-effective platform for hosting large-scale service appli- cations. While large data centers enjoy economies of scale by amortizing initial capital investment over large number of machines, they also incur tremendous energy cost in terms of power distribution and cooling. An effective approach for saving energy in data centers is to adjust dynamically the data center capacity by turning off unused machines. How- ever, this dynamic capacity provisioning problem is known to be challenging as it requires a careful understanding of the resource demand characteristics as well as considerations to various cost factors, including task scheduling delay, ma- chine reconfiguration cost and electricity price fluctuation. In this paper, we provide a control-theoretic solution to the dynamic capacity provisioning problem that minimizes the total energy cost while meeting the performance objec- tive in terms of task scheduling delay. Specifically, we model this problem as a constrained discrete-time optimal control problem, and use Model Predictive Control (MPC) to find the optimal control policy. Through extensive analysis and simulation using real workload traces from Google's compute clusters, we show that our proposed framework can achieve significant reduction in energy cost, while maintaining an acceptable average scheduling delay for individual tasks.

AB - Data centers have recently gained significant popularity as a cost-effective platform for hosting large-scale service appli- cations. While large data centers enjoy economies of scale by amortizing initial capital investment over large number of machines, they also incur tremendous energy cost in terms of power distribution and cooling. An effective approach for saving energy in data centers is to adjust dynamically the data center capacity by turning off unused machines. How- ever, this dynamic capacity provisioning problem is known to be challenging as it requires a careful understanding of the resource demand characteristics as well as considerations to various cost factors, including task scheduling delay, ma- chine reconfiguration cost and electricity price fluctuation. In this paper, we provide a control-theoretic solution to the dynamic capacity provisioning problem that minimizes the total energy cost while meeting the performance objec- tive in terms of task scheduling delay. Specifically, we model this problem as a constrained discrete-time optimal control problem, and use Model Predictive Control (MPC) to find the optimal control policy. Through extensive analysis and simulation using real workload traces from Google's compute clusters, we show that our proposed framework can achieve significant reduction in energy cost, while maintaining an acceptable average scheduling delay for individual tasks.

KW - Cloud computing

KW - Energy manage- ment

KW - Model predictive control

KW - Resource management

UR - http://www.scopus.com/inward/record.url?scp=84867698582&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867698582&partnerID=8YFLogxK

U2 - 10.1145/2371536.2371562

DO - 10.1145/2371536.2371562

M3 - Conference contribution

SN - 9781450315203

SP - 145

EP - 154

BT - ICAC'12 - Proceedings of the 9th ACM International Conference on Autonomic Computing

ER -