An Adaptive and Efficient Design and Implementation for Meteorology Data Grid Using Grid Technology

Grid technologies not only provide the infrastructure to collaborate among virtual organizations to share heterogeneous resources but also provide a service oriented paradigm for business and social software development. In this paper, we present an agent based efficient and adaptive design for our meteorological data portal system, called meteo-data grid and its implementation using grid technology. Our adaptive agent based approach analyzes the workload and then invoke or revoke agents using factory design pattern to process the data files of gigabyte scale. We also investigated the suitability of grid technologies for a portal based data management system in comparison with currently available Web technologies, namely Web services and Java servlets and evaluated their performance in terms of response time to retrieve varying amount of data from database and serve to GridSphere portlet. Our experiments results show that to serve large scale of data, grid services are much faster as compared to Web services and Java servlets