A Replica Replacement Algorithm Based on Value-Cost Prediction

  • Authors:
  • Tian Tian;Junzhou Luo;Zhiang Wu

  • Affiliations:
  • School of Computer Science and Engineering, Southeast University, Nanjing, P.R. China 210096;School of Computer Science and Engineering, Southeast University, Nanjing, P.R. China 210096;School of Computer Science and Engineering, Southeast University, Nanjing, P.R. China 210096

  • Venue:
  • Computer Supported Cooperative Work in Design IV
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Data Grid is a new kind of data management infrastructure. To improve access efficiency, large amounts of data on Data Grids need to be replicated in the form of replicas at distributed sites. Due to the limited storage capacity, a good replica replacement algorithm is needed. This paper proposes a replica replacement algorithm based on value-cost prediction, which achieves a good balance between value and cost. The algorithm predicts replica value to select replicas that will be replaced in the first stage, and predicts the bandwidth to make replacement cost as low as possible in the second stage. Simulation shows that compared with traditional replacement algorithms our replica replacement algorithm contributes to better grid performance.