Presentation 2011-03-03
Machine Learning Based Power Management for WWW Server Clusters
Takehito HIROTA, Satoru OHTA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The WWW (World Wide Web) service is often provided by a server cluster, which consists of multiple server computers. It is significant to save power consumed in a server cluster from the viewpoint of operational expenditure as well as the global environment. The power of a cluster is saved by measuring the load on the server and operating as few computers as possible. For this operation, it is necessary to measure several load metrics and decide the required number of server computers, with considering the relationship among the metrics. As a method of deciding the number of server computers, this paper proposes a scheme based on the machine learning technique. The paper first presents how the technique is applied to the decision of the computer number. Then, the implementation is presented. The effectiveness of the scheme is confirmed through experiments.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) machine learning / server cluster / power management / traffic
Paper # IN2010-177
Date of Issue

Conference Information
Committee IN
Conference Date 2011/2/24(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Information Networks (IN)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Machine Learning Based Power Management for WWW Server Clusters
Sub Title (in English)
Keyword(1) machine learning
Keyword(2) server cluster
Keyword(3) power management
Keyword(4) traffic
1st Author's Name Takehito HIROTA
1st Author's Affiliation Faculty of Engineering, Toyama Prefectural University()
2nd Author's Name Satoru OHTA
2nd Author's Affiliation Faculty of Engineering, Toyama Prefectural University
Date 2011-03-03
Paper # IN2010-177
Volume (vol) vol.110
Number (no) 449
Page pp.pp.-
#Pages 6
Date of Issue