Optimization of online disk scheduling algorithms

  • Authors:
  • Saman Zarandioon;Alexander Thomasian

  • Affiliations:
  • New Jersey Institute of Technology, Newark, NJ;New Jersey Institute of Technology, Newark, NJ

  • Venue:
  • ACM SIGMETRICS Performance Evaluation Review - Design, implementation, and performance of storage systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Disk scheduling policies have a significant effect on disk performance. SPTF is one of the well-known policies that can increase disk performance near to optimality [1]. One of the drawbacks of the regular implementation of SPTF is its high computational cost. "The computational cost [of SPTF] (as indicated crudely by our simulation times) is very high" [2]. This paper shows that computational cost of SPTF is not the characteristic of SPTF, but it is a matter of implementation. The experience shows that this approach can improve the efficiency over 80% compared to naïve implementation. Finally, an algorithm for efficient implementation of lookahead algorithms is introduced.