Modeling and Performance Evaluation of Hybrid Storage I/O in Data Grid

  • Authors:
  • Zhaobin Liu;Haitao Li

  • Affiliations:
  • Dalian Maritime University;Dalian Maritime University

  • Venue:
  • NPC '07 Proceedings of the 2007 IFIP International Conference on Network and Parallel Computing Workshops
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Storage I/O has become a performance bottleneck for many Data Grid applications. Characterization of I/O patterns has shown that many of these applications have complex, irregular I/O patterns. In this paper we show how Stochastic Petri Net (SPN) models can be exploited for performance analysis of hybrid I/O Data Grid storage systems. We study a typical storage system SPN modeling and simplify model complexities based on aggregate I/O. Evaluation using case studies shows that we can adjust the priority schedule by changing the ratio of file I/O and multimedia I/O.