Data Replication Using Experience Based Trust in a Data Grid Environment

High availability is the desirable feature of the data in a data grid. The data is replicated in a data grid to increase the data availability, reliability and improve the response time of the applications by reducing the network traffic. The efficiency and fast access to such replicated data are influenced by the location of the resource holding the replica and the access cost, The replica locations may be some times unreliable and there is a risk that the resources may fail and hence fail to provide the guaranteed QoS. The notion of trust; is to manage such risk and improve the QoS provided by the resource locations. In this paper we apply the concept of experience based trust to the problem of data replication in a data grid environment,. We proposed different trust-based replica location selection algorithms and evaluated their effectiveness for a simulated grid. The aim of this work is to identify and to acquire the most available, least loaded and the fastest, resource is a replica location.