A network model for prediction of temperature distribution in data centers

Shinya Tashiro, Yuya Tarutani, Go Hasegawa, Yutaka Nakamura, Kazuhiro Matsuda, Morito Matsuoka

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

6 Citations (Scopus)

Abstract

We propose a novel network model for real-time prediction of temperature distribution in a data center so as to allow energy-efficient task assignment and facility management. We model various physical relationships in the data center as a network, including heat movements caused by airflow and heat generation by servers. Since changes in temperature distribution depend on physical properties of the data center such as equipment locations and server types, model parameters (connection weights in the network) that characterize relationship of nodes are determined by a machine learning technique using actual data center operation data. The proposed method provides prediction results in a shorter time than traditional methods such as model based on computational fluid dynamics and potential flow model, while maintaining prediction accuracy. We evaluate the performance of the proposed model through comparison with actual data from our experimental data center. The evaluation indicates that the proposed model can predict 10-minute future temperature distributions in 60 places in 3.3 ms, with a root mean square error of 0.49 degrees.

Original languageEnglish
Title of host publication2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages261-266
Number of pages6
ISBN (Electronic)9781467395014
DOIs
Publication statusPublished - Nov 20 2015
Externally publishedYes
Event4th IEEE International Conference on Cloud Networking, CloudNet 2015 - Falls, Canada
Duration: Oct 5 2015Oct 7 2015

Publication series

Name2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015

Conference

Conference4th IEEE International Conference on Cloud Networking, CloudNet 2015
CountryCanada
CityFalls
Period10/5/1510/7/15

Fingerprint

Temperature distribution
Servers
Potential flow
Heat generation
Mean square error
Learning systems
Computational fluid dynamics
Physical properties

Keywords

  • Data center
  • Machine learning
  • Network model
  • Power consumption reduction
  • Temperature prediction

ASJC Scopus subject areas

  • Computer Networks and Communications

Cite this

Tashiro, S., Tarutani, Y., Hasegawa, G., Nakamura, Y., Matsuda, K., & Matsuoka, M. (2015). A network model for prediction of temperature distribution in data centers. In 2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015 (pp. 261-266). [7335319] (2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CloudNet.2015.7335319

A network model for prediction of temperature distribution in data centers. / Tashiro, Shinya; Tarutani, Yuya; Hasegawa, Go; Nakamura, Yutaka; Matsuda, Kazuhiro; Matsuoka, Morito.

2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015. Institute of Electrical and Electronics Engineers Inc., 2015. p. 261-266 7335319 (2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015).

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

Tashiro, S, Tarutani, Y, Hasegawa, G, Nakamura, Y, Matsuda, K & Matsuoka, M 2015, A network model for prediction of temperature distribution in data centers. in 2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015., 7335319, 2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015, Institute of Electrical and Electronics Engineers Inc., pp. 261-266, 4th IEEE International Conference on Cloud Networking, CloudNet 2015, Falls, Canada, 10/5/15. https://doi.org/10.1109/CloudNet.2015.7335319
Tashiro S, Tarutani Y, Hasegawa G, Nakamura Y, Matsuda K, Matsuoka M. A network model for prediction of temperature distribution in data centers. In 2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015. Institute of Electrical and Electronics Engineers Inc. 2015. p. 261-266. 7335319. (2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015). https://doi.org/10.1109/CloudNet.2015.7335319
Tashiro, Shinya ; Tarutani, Yuya ; Hasegawa, Go ; Nakamura, Yutaka ; Matsuda, Kazuhiro ; Matsuoka, Morito. / A network model for prediction of temperature distribution in data centers. 2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015. Institute of Electrical and Electronics Engineers Inc., 2015. pp. 261-266 (2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015).
@inproceedings{086f94d6622d4495acc1f8f90fdf3641,
title = "A network model for prediction of temperature distribution in data centers",
abstract = "We propose a novel network model for real-time prediction of temperature distribution in a data center so as to allow energy-efficient task assignment and facility management. We model various physical relationships in the data center as a network, including heat movements caused by airflow and heat generation by servers. Since changes in temperature distribution depend on physical properties of the data center such as equipment locations and server types, model parameters (connection weights in the network) that characterize relationship of nodes are determined by a machine learning technique using actual data center operation data. The proposed method provides prediction results in a shorter time than traditional methods such as model based on computational fluid dynamics and potential flow model, while maintaining prediction accuracy. We evaluate the performance of the proposed model through comparison with actual data from our experimental data center. The evaluation indicates that the proposed model can predict 10-minute future temperature distributions in 60 places in 3.3 ms, with a root mean square error of 0.49 degrees.",
keywords = "Data center, Machine learning, Network model, Power consumption reduction, Temperature prediction",
author = "Shinya Tashiro and Yuya Tarutani and Go Hasegawa and Yutaka Nakamura and Kazuhiro Matsuda and Morito Matsuoka",
year = "2015",
month = "11",
day = "20",
doi = "10.1109/CloudNet.2015.7335319",
language = "English",
series = "2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "261--266",
booktitle = "2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015",

}

TY - GEN

T1 - A network model for prediction of temperature distribution in data centers

AU - Tashiro, Shinya

AU - Tarutani, Yuya

AU - Hasegawa, Go

AU - Nakamura, Yutaka

AU - Matsuda, Kazuhiro

AU - Matsuoka, Morito

PY - 2015/11/20

Y1 - 2015/11/20

N2 - We propose a novel network model for real-time prediction of temperature distribution in a data center so as to allow energy-efficient task assignment and facility management. We model various physical relationships in the data center as a network, including heat movements caused by airflow and heat generation by servers. Since changes in temperature distribution depend on physical properties of the data center such as equipment locations and server types, model parameters (connection weights in the network) that characterize relationship of nodes are determined by a machine learning technique using actual data center operation data. The proposed method provides prediction results in a shorter time than traditional methods such as model based on computational fluid dynamics and potential flow model, while maintaining prediction accuracy. We evaluate the performance of the proposed model through comparison with actual data from our experimental data center. The evaluation indicates that the proposed model can predict 10-minute future temperature distributions in 60 places in 3.3 ms, with a root mean square error of 0.49 degrees.

AB - We propose a novel network model for real-time prediction of temperature distribution in a data center so as to allow energy-efficient task assignment and facility management. We model various physical relationships in the data center as a network, including heat movements caused by airflow and heat generation by servers. Since changes in temperature distribution depend on physical properties of the data center such as equipment locations and server types, model parameters (connection weights in the network) that characterize relationship of nodes are determined by a machine learning technique using actual data center operation data. The proposed method provides prediction results in a shorter time than traditional methods such as model based on computational fluid dynamics and potential flow model, while maintaining prediction accuracy. We evaluate the performance of the proposed model through comparison with actual data from our experimental data center. The evaluation indicates that the proposed model can predict 10-minute future temperature distributions in 60 places in 3.3 ms, with a root mean square error of 0.49 degrees.

KW - Data center

KW - Machine learning

KW - Network model

KW - Power consumption reduction

KW - Temperature prediction

UR - http://www.scopus.com/inward/record.url?scp=84960955614&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84960955614&partnerID=8YFLogxK

U2 - 10.1109/CloudNet.2015.7335319

DO - 10.1109/CloudNet.2015.7335319

M3 - Conference contribution

AN - SCOPUS:84960955614

T3 - 2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015

SP - 261

EP - 266

BT - 2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015

PB - Institute of Electrical and Electronics Engineers Inc.

ER -