Architecture of Geospatial Big-Data Batch Processing Model Based on Hadoop

In this paper, we propose a spatial ETL toolkit model, regardless of collectible data format, to replace traditional ETL (Extract, Transform, Load) to process public or individual system server with fast data creation. Recently various web servers, and the need of analyzing structured/unstructured Big-Data log information generated by independent high-speed servers are increasing. In addition, the space ETL model for replacing an existing service, which is built into a relational database is required and replaced the existing system in order to analyze unstructured data log information in Big-Data environments, where space is needed the ETL model. This paper proposes the spatial ETL model and proposed model is possible loading log information regardless of connected Hadoop or local storage to input data format quickly and accurately. Existing relational database consist of structured data is possible connection and extend function.