Thermal influence indices: Causality metrics for efficient exploration of data center cooling

Harshad Bhagwat, Amarendra Singh, Arunchandar Vasan, Anand Sivasubramaniam

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

11 Citations (Scopus)

Abstract

Cooling is an important issue in data center design and operation. Accurate evaluation of a design or operational parameter choice for cooling is difficult as it requires several runs of computationally intensive Computational Fluid Dynamics (CFD) based models. Therefore there is need for an exploration method that does not incur enormous computation. In addition, the exploration should also provide insights that enable informed decision making. Given these twin goals of reduced computation and improved insights, we present a novel approach to data center cooling exploration. The key idea is to do a local search around the current design/operation of a data center to obtain better design/operation parameters subject to the desired constraints. To do this, all the microscopic information about airflow and temperature in data center available from a single run of CFD computation is converted into macroscopic metrics called influence indices. The influence indices, which characterize the causal relationship between heat sources and sinks, are used to refine the design/operation of the data center either manually or programmatically. New designs are evaluated with further CFD runs to compute new influence indices and the process is repeated to yield improved designs as per the computation budget available. We have carried out design exploration of a realistic data center using this methodology. Specifically, we considered maximization of the heat load in the data center subject to the constraints that: 1) servers are kept at appropriate temperatures and 2) overloading of CRACs is avoided. Our evaluation shows that the use of influence indices cuts down the exploration time by 80 % for a 1500 sq. ft. data center.

Original languageEnglish (US)
Title of host publication2012 International Green Computing Conference, IGCC 2012
DOIs
StatePublished - Nov 26 2012
Event2012 International Green Computing Conference, IGCC 2012 - San Jose, CA, United States
Duration: Jun 4 2012Jun 8 2012

Publication series

Name2012 International Green Computing Conference, IGCC 2012

Other

Other2012 International Green Computing Conference, IGCC 2012
CountryUnited States
CitySan Jose, CA
Period6/4/126/8/12

Fingerprint

Cooling
Computational fluid dynamics
Hot Temperature
Thermal load
Servers
Decision making
Temperature

All Science Journal Classification (ASJC) codes

  • Software

Cite this

Bhagwat, H., Singh, A., Vasan, A., & Sivasubramaniam, A. (2012). Thermal influence indices: Causality metrics for efficient exploration of data center cooling. In 2012 International Green Computing Conference, IGCC 2012 [6322254] (2012 International Green Computing Conference, IGCC 2012). https://doi.org/10.1109/IGCC.2012.6322254
Bhagwat, Harshad ; Singh, Amarendra ; Vasan, Arunchandar ; Sivasubramaniam, Anand. / Thermal influence indices : Causality metrics for efficient exploration of data center cooling. 2012 International Green Computing Conference, IGCC 2012. 2012. (2012 International Green Computing Conference, IGCC 2012).
@inproceedings{60127293c1a042bd89b229a9e49dd452,
title = "Thermal influence indices: Causality metrics for efficient exploration of data center cooling",
abstract = "Cooling is an important issue in data center design and operation. Accurate evaluation of a design or operational parameter choice for cooling is difficult as it requires several runs of computationally intensive Computational Fluid Dynamics (CFD) based models. Therefore there is need for an exploration method that does not incur enormous computation. In addition, the exploration should also provide insights that enable informed decision making. Given these twin goals of reduced computation and improved insights, we present a novel approach to data center cooling exploration. The key idea is to do a local search around the current design/operation of a data center to obtain better design/operation parameters subject to the desired constraints. To do this, all the microscopic information about airflow and temperature in data center available from a single run of CFD computation is converted into macroscopic metrics called influence indices. The influence indices, which characterize the causal relationship between heat sources and sinks, are used to refine the design/operation of the data center either manually or programmatically. New designs are evaluated with further CFD runs to compute new influence indices and the process is repeated to yield improved designs as per the computation budget available. We have carried out design exploration of a realistic data center using this methodology. Specifically, we considered maximization of the heat load in the data center subject to the constraints that: 1) servers are kept at appropriate temperatures and 2) overloading of CRACs is avoided. Our evaluation shows that the use of influence indices cuts down the exploration time by 80 {\%} for a 1500 sq. ft. data center.",
author = "Harshad Bhagwat and Amarendra Singh and Arunchandar Vasan and Anand Sivasubramaniam",
year = "2012",
month = "11",
day = "26",
doi = "10.1109/IGCC.2012.6322254",
language = "English (US)",
isbn = "9781467321556",
series = "2012 International Green Computing Conference, IGCC 2012",
booktitle = "2012 International Green Computing Conference, IGCC 2012",

}

Bhagwat, H, Singh, A, Vasan, A & Sivasubramaniam, A 2012, Thermal influence indices: Causality metrics for efficient exploration of data center cooling. in 2012 International Green Computing Conference, IGCC 2012., 6322254, 2012 International Green Computing Conference, IGCC 2012, 2012 International Green Computing Conference, IGCC 2012, San Jose, CA, United States, 6/4/12. https://doi.org/10.1109/IGCC.2012.6322254

Thermal influence indices : Causality metrics for efficient exploration of data center cooling. / Bhagwat, Harshad; Singh, Amarendra; Vasan, Arunchandar; Sivasubramaniam, Anand.

2012 International Green Computing Conference, IGCC 2012. 2012. 6322254 (2012 International Green Computing Conference, IGCC 2012).

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

TY - GEN

T1 - Thermal influence indices

T2 - Causality metrics for efficient exploration of data center cooling

AU - Bhagwat, Harshad

AU - Singh, Amarendra

AU - Vasan, Arunchandar

AU - Sivasubramaniam, Anand

PY - 2012/11/26

Y1 - 2012/11/26

N2 - Cooling is an important issue in data center design and operation. Accurate evaluation of a design or operational parameter choice for cooling is difficult as it requires several runs of computationally intensive Computational Fluid Dynamics (CFD) based models. Therefore there is need for an exploration method that does not incur enormous computation. In addition, the exploration should also provide insights that enable informed decision making. Given these twin goals of reduced computation and improved insights, we present a novel approach to data center cooling exploration. The key idea is to do a local search around the current design/operation of a data center to obtain better design/operation parameters subject to the desired constraints. To do this, all the microscopic information about airflow and temperature in data center available from a single run of CFD computation is converted into macroscopic metrics called influence indices. The influence indices, which characterize the causal relationship between heat sources and sinks, are used to refine the design/operation of the data center either manually or programmatically. New designs are evaluated with further CFD runs to compute new influence indices and the process is repeated to yield improved designs as per the computation budget available. We have carried out design exploration of a realistic data center using this methodology. Specifically, we considered maximization of the heat load in the data center subject to the constraints that: 1) servers are kept at appropriate temperatures and 2) overloading of CRACs is avoided. Our evaluation shows that the use of influence indices cuts down the exploration time by 80 % for a 1500 sq. ft. data center.

AB - Cooling is an important issue in data center design and operation. Accurate evaluation of a design or operational parameter choice for cooling is difficult as it requires several runs of computationally intensive Computational Fluid Dynamics (CFD) based models. Therefore there is need for an exploration method that does not incur enormous computation. In addition, the exploration should also provide insights that enable informed decision making. Given these twin goals of reduced computation and improved insights, we present a novel approach to data center cooling exploration. The key idea is to do a local search around the current design/operation of a data center to obtain better design/operation parameters subject to the desired constraints. To do this, all the microscopic information about airflow and temperature in data center available from a single run of CFD computation is converted into macroscopic metrics called influence indices. The influence indices, which characterize the causal relationship between heat sources and sinks, are used to refine the design/operation of the data center either manually or programmatically. New designs are evaluated with further CFD runs to compute new influence indices and the process is repeated to yield improved designs as per the computation budget available. We have carried out design exploration of a realistic data center using this methodology. Specifically, we considered maximization of the heat load in the data center subject to the constraints that: 1) servers are kept at appropriate temperatures and 2) overloading of CRACs is avoided. Our evaluation shows that the use of influence indices cuts down the exploration time by 80 % for a 1500 sq. ft. data center.

UR - http://www.scopus.com/inward/record.url?scp=84869447292&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84869447292&partnerID=8YFLogxK

U2 - 10.1109/IGCC.2012.6322254

DO - 10.1109/IGCC.2012.6322254

M3 - Conference contribution

AN - SCOPUS:84869447292

SN - 9781467321556

T3 - 2012 International Green Computing Conference, IGCC 2012

BT - 2012 International Green Computing Conference, IGCC 2012

ER -

Bhagwat H, Singh A, Vasan A, Sivasubramaniam A. Thermal influence indices: Causality metrics for efficient exploration of data center cooling. In 2012 International Green Computing Conference, IGCC 2012. 2012. 6322254. (2012 International Green Computing Conference, IGCC 2012). https://doi.org/10.1109/IGCC.2012.6322254