Imbalance of CPU temperatures in a blade system and its impact for power consumption of fans

Authors:
Yuetsu Kodama;Satoshi Itoh;Toshiyuki Shimizu;Satoshi Sekiguchi;Hiroshi Nakamura;Naohiko Mori
Affiliations:
Graduate School of System and Information Engineering, University of Tsukuba, Tsukuba, Japan;Information Technology Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan;Information Technology Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan;Information Technology Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan;Department of Information Physics and Computing, The University of Tokyo, Tokyo, Japan;Applied Network Integration Business Unit, NTT Advanced Technology Corporation, Tokyo, Japan and Innovative IP Architecture Center, NTT Communications Corporation, Tokyo, Japan
Venue:
Cluster Computing
Year:
2013

Citing 4
Cited 0

Managing energy and server resources in hosting centers

SOSP '01 Proceedings of the eighteenth ACM symposium on Operating systems principles
Power Aware Scheduling of Bag-of-Tasks Applications with Deadline Constraints on DVS-enabled Clusters

CCGRID '07 Proceedings of the Seventh IEEE International Symposium on Cluster Computing and the Grid
Joint optimization of idle and cooling power in data centers while maintaining response time

Proceedings of the fifteenth edition of ASPLOS on Architectural support for programming languages and operating systems
Demystifying energy consumption in Grids and Clouds

GREENCOMP '10 Proceedings of the International Conference on Green Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

We are now developing a new metric of data center power efficiency to fairly evaluate the contribution of each improvement for power efficiency. In order to develop it, we built a testbed of a data center and measured power consumption of each components and environmental variables in some detail, including the power consumption and temperature of each node, rack and air conditioning unit, as well as load on the CPU, Disk I/O and the network. In these measurements we found that there was a significant imbalance of CPU temperatures that caused an imbalance in the power consumption of fans. We clarified the relationship between CPU load and fan speed, and showed that scheduling or rearrangement of nodes could reduce the power consumption of fans. We reduced fan power consumption by a maximum of 62% and total power consumption by a maximum of 12% by changing the scheduling of five nodes, changing the nodes used from hot nodes to cool nodes.