Chameleon: Adapting throughput server to time-varying green power budget using online learning

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Eco-friendly energy sources (i.e. green power) attract great attention as lowering computer carbon footprint has become a necessity. Existing proposals on managing green energy powered systems show sub-optimal results since they either use rigid load power capping or heavily rely on backup power. We propose Chameleon, a novel adaptive green throughput server. Chameleon comprises of multiple flexible power management policies and leverages learning algorithm to select the optimal operating mode during runtime. The proposed design outperforms the state-of-the-art approach by 13% on performance, improves system MTBF by 42%, and still maintains up to 95% green energy utilization.

Original languageEnglish
Title of host publicationProceedings of the International Symposium on Low Power Electronics and Design, ISLPED 2013
Pages100-105
Number of pages6
DOIs
StatePublished - 2013
Event2013 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED 2013 - Beijing, China
Duration: 4 Sep 20136 Sep 2013

Publication series

NameProceedings of the International Symposium on Low Power Electronics and Design
ISSN (Print)1533-4678

Conference

Conference2013 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED 2013
Country/TerritoryChina
CityBeijing
Period4/09/136/09/13

Keywords

  • adaptation
  • green power
  • learning
  • throughput server

Fingerprint

Dive into the research topics of 'Chameleon: Adapting throughput server to time-varying green power budget using online learning'. Together they form a unique fingerprint.

Cite this