Optimizing Storage System Power and Performance

Open Access
Garg, Rajat
Graduate Program:
Computer Science and Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
April 11, 2011
Committee Members:
  • Mahmut Taylan Kandemir, Thesis Advisor
  • partitioning
  • control
  • markov
  • cache
  • idle time
The quest of building bigger and better computing systems has resulted in tremendous growth in the size of the storage systems. Not only have they grown in their size, they play a significant role in determining the overall performance of the applications and success of the entire computing system. While the industry is concerned about reducing the huge costs involved in running/maintaining these storage systems, the scientific community has been pushing the limit to achieve maximum performance. Apart from the contrasting demands of saving power vs. maximum performance, there exist scenarios where a balance of power consumption and performance is expected. In this body of work, we propose/study software based techniques that will help achieve some or all of the above requirements.