HoTMaP: Holistic Thermal Management and Power Capping for Green Data Centers

Project Overview


To accommodate modern resource-intensive high-performance applications, large-scale computing and storage platforms have grown at a rapid pace in a variety of domains ranging from research labs and academic groups to industry. The fast-growing power consumption of these platforms is a major concern due to its implications on cost and efficiency of these platforms as well as the well-being of our environment. Trends from such platforms suggest that the power consumption in high-performance computing platforms (or data centers) already accounted for 1.2% of the overall electricity consumption in the country in 2005. More alarmingly, if current practices for the design and operation of these platforms continue, their power consumption is projected to keep growing at 18% every year. Related to this, operating the hardware within a data center within safe thermal regimes is also an important concern to ensure reliability.
This project is concerned with the following two meta-issues: (i) Goals: How to determine desirable performance/power/thermal operating regimes in state-of-the art data centers? (ii) Mechanisms: How to operate data centers in these regimes?

People


Jeong Hwan Choi
Sriram Govindan
Anand Sivasubramaniam
Bhuvan Urgaonkar
Di Wang

 

Publications

Copyright notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author’s copyright. These works may not be reposted without the explicit permission of the copyright holder.

Energy Storage in Datacenters: What, Where and How Much?
Di Wang, Chuangang Ren, Anand Sivasubramaniam, Bhuvan Urgaonkar, and Hosam K. Fathy
Proceedings of the ACM International Conference on Measurement and Modeling of Computer Systems (SIGMETRICS 2012), London, UK. June 2012.
Paper

Leveraging Stored Energy for Handling Power Emergencies in Aggressively Provisioned Datacenters
Sriram Govindan, Di Wang, Anand Sivasubramaniam, and Bhuvan Urgaonkar
Proceedings of the International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS 2012), London, UK. March 2012.
Paper

Benefits and Limitations of Tapping into Stored Energy for Datacenters
Sriram Govindan, Anand Sivasubramaniam, and Bhuvan Urgaonkar
Proceedings of the International Symposium on Computer Architecture (ISCA), San Jose, CA, June 2011. (40/208 = 19% accepted)
Paper

Optimal Power Cost Management Using Stored Energy in Data Centers
Rahul Urgaonkar, Bhuvan Urgaonkar, Michael Neely, and Anand Sivasubramaniam
Proceedings of the ACM International Conference on Measurement and Modeling of Computer Systems (SIGMETRICS), San Jose, CA, June 2011. (26/177 = 14% accepted)
Paper

Statistical Profiling-based Techniques for Effective Power Provisioning in Data Centers
Sriram Govindan, Jeonghwan Choi, Bhuvan Urgaonkar, Anand Sivasubramaniam, and Andrea Baldini
Proceedings of EuroSys, sponsored by the European Chapter of SIGOPS (EUROSYS), April 2009, Munich, Germany. (25/148 = 17% accepted).
Paper

Empirical Analysis of the Energy Efficiency of Flash-based SSDs
Euiseong Seo, Seonyeong Park, and Bhuvan Urgaonkar
Proceedings of the First Workshop on Power Aware Computing and Systems (HotPower), held in conjunction with USENIX OSDI 2008, December 2008, San Diego, CA.
Paper

Profiling, Prediction, and Capping of Power Consumption for Consolidated Environments
Jeonghwan Choi, Sriram Govindan, Bhuvan Urgaonkar, and Anand Sivasubramanian
Proceedings of the Sixteenth Annual Meeting of the IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS), September 2008, Baltimore, MD. (36/94 = 38% accepted). Best student paper award.
Technical Report CSE 08-006, Dept. of Computer Science and Engg., The Pennsylvania State University, March 2008.
Paper    Tech. Report

QDSL: A Queueing Model for Systems with Differential Service Levels
Shiva Chaitanya, Bhuvan Urgaonkar, and Anand Sivasubramaniam
Proceedings of the ACM International Conference on Measurement and Modeling of Computer Systems (SIGMETRICS), pages 289-300, Annapolis, MD, June 2008. (36/201 = 18% accepted).
Paper

CFD-based Tool for Studying Temperature in Rack-mounted Servers
J. Choi, Y. Kim, A. Sivasubramaniam, J. Srebric, Q. Wang, and J. Lee
IEEE Transactions on Computer (TC), IEEE Press, August 2008.
Paper

Modeling and Managing Thermal Profiles of Rack-mounted Servers with ThermoStat
J. Choi, Y. Kim, A. Sivasubramaniam, J. Srebric, Q. Wang, and J. Lee
Proceedings of the International Symposium on High Performance Computer Architecture (HPCA), February 2007. (28/174 = 16% accepted). Best student paper candidate.
Paper

 

Software


Aggregate Power Prediction: Scripts for predicting the power usage of application aggregates.
Watermark-based Power Budget Enforcement: Distributed code to enforce specified sustained power budget for a server.
Thermostat V0.9: Thermal modeling tool for rack-mounted servers.

 

Power Measurement Infrastructure


High Precision Multimeter
PDU Infrastructure

 

Sponsors


We are grateful to the following for supporting our research:

 

  • Google, Inc.
  • HP