Utility-Based Replication Strategies in Data Grids
Authors:
- 1 January 2006
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 500-507
Abstract
Providing fast, reliable and transparent access to data to all users within a community is one of the most crucial functions of data management in a grid environment. Replication strategies are regarded as one of the major optimization techniques for reducing access latency, improving data locality, and increasing robustness, scalability, and performance for the data grids. Compared with economy-based strategies implemented in OptorSim, utility-based strategies proposed in this paper, which make "buying" or replacement decision based on the utility model, achieve better performance in job execution cost and storage consumption with the same jobs being finished in three different type data grids. The simulation experiments also show that the larger the test data grid scale is, the more the job execution cost saved, which makes our method have a greater advantage in continuously growing data gridsKeywords
This publication has 8 references indexed in Scilit:
- A Peer-to-Peer Replica Location Service Based on a Distributed Hash Table2005
- Replica Management in the European DataGrid ProjectJournal of Grid Computing, 2004
- Simulation of dynamic data replication strategies in Data Grids2004
- Analysis of Scheduling and Replica Optimisation Strategies for Data Grids Using OptorSimJournal of Grid Computing, 2004
- Replica Location Mechanism Based on DHT and the Small-World TheoryLecture Notes in Computer Science, 2004
- Towards an Economy-Based Optimisation of File Access and Replication on a Data Grid2003
- Improving Data Availability through Dynamic Model-Driven Replication in Large Peer-to-Peer Communities2003
- The data grid: Towards an architecture for the distributed management and analysis of large scientific datasetsJournal of Network and Computer Applications, 2000