An Intelligent ETL Grid-Based Solution to Enable Spatial Data Warehouse Deployment in Cyber Physical System Context

Thanks to their spatially distributed sensors, Cyber-Physical System (CPS) applications are currently accumulating large amounts of heterogeneous data. When it comes to allowing several decision-makers to collaboratively plan their actions, these applications need appropriate tools for an efficient storage, analysis, and visualization of the available data. Spatial Data Warehouses (SDWs) have proven their efficiency in carrying out these operations. However, because of the increasing quantity of data, the Extract-Transform-Load (ETL) process (which is in charge of aggregating several data sources within a unified data storage repository) generally fails to update the SDW within predefined window times. In order to solve this problem, we propose in this paper to distribute the ETL tasks over a grid of computing resources. We also propose a multiagent-based approach that controls the ETL grid while allowing a convenient use of the shared resources. In addition to being the unique solution that uses grid computing for the ETL process of SDWs, our approach allows a joint use of archive and real-time data for personalized reporting and visualization of services envisioned to the decision-makers who are using the CPS application