Power-aware DBMS: potential and challenges

  • Authors:
  • Yi-Cheng Tu;Xiaorui Wang;Zichen Xu

  • Affiliations:
  • Department of Computer Science and Engineering, Tampa, FL;Department of Electrical Engineering and Computer Science, The University of Tennessee, Knoxville, TN;Department of Computer Science and Engineering, Tampa, FL

  • Venue:
  • SSDBM'11 Proceedings of the 23rd international conference on Scientific and statistical database management
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Energy consumption has become a first-class optimization goal in computing system design and implementation. Database systems, being a major consumer of computing resources (thus energy) in modern data centers, also face the challenges of going green. In this position paper, we describe our vision on this new direction of database system research, and report the results of our recent work on this topic. We describe our ideas on the key issues in designing a power-aware DBMS and sketch our solutions to such issues. Specifically, we believe that the ability for the DBMS to dynamically adjust various knobs to satisfy energy (and performance) goals is the main technical challenge in this paradigm. To address that challenge, we propose dynamic modeling and tuning techniques based on formal feedback control theory. Our preliminary data clearly show that the energy savings can be significant.