An Innovative Approach of Data Grid Consistency Using Tree-Based Clustering

Abstract

Nowadays, Data Grids can be seen as a frameworks responding to the needs of large scale applications by affording varied geographically distributed resources set. The main aim here is to ensure flexible and efficient access and quality data, to improve the availability and accessibility, we must improve scalability and efficient usability of Grid to reduce application communication costs and request average response times. In relation with the problem of the coherence maintenance of replicas, we propose a Tree-Based Clustering method which not only organizes those replicas having similarity in demand number into clusters hierarchically looking like trees, but also tries to choose the shortest connection so that communication cost, average response time and total data transferred through Grid will be reduced.