The numbers of source system and the amount of data requiring manipulation have increased significantly. As a result, analyzing characteristics of data and making statistics have become more difficult and complex. This paper improved the efficiency of data analysis and statistics by utilizing a BI(Business Intelligence) system in the following ways. First, using an ETL(Extract, Transform, Loading) solution like SSIS(SQL Server Integration Service), the data in the view tables of source systems was extracted and transformed before it was loaded into the ODS(Operational Data Store) within the BI system. Then, the data in ODS was cleansed and loaded, before being stored in DW (Data Warehouse).
Then, the SSAS(SQL Server Analysis Service) solution was used for analyzing data and creating cube DB. Utilizing the SSAS solution, the data in the DW was divided into small DB that is called as cube DB.
As the way of creating report, the SSRS(SQL Server Reporting Services) solution provided reporting service and graphical reports were developed by PPS(Performance Point Service). Graphic-based reports can provide drill-down and drill-up, pivot function for multi-dimensional data analysis.
In addition to the previously mentioned benefits of the BI system, this paper produced the following beneficial results. First, the per subject view table made in the source system allows the data to be created in real time. Then, the data was accurately extracted and collected from view table into ODS. Second, the BI system allows users to generate reports by extracting data from the cube DB directly. Third, users can analyze report using multi-dimensional methods like drill-down and drill-up, pivot function.
In terms of performance and functionality, using the cube DB provides faster performance than using source systems DB. Also, in comparison to source systems, the BI system provides more powerful statistics and data analysis
However, the deficiency of BI system is poor performance as the kinds of source system and the complexity of business increase. In order to compensate this deficiency, the framework based on Hadoop that includes HDFS(High Distributed File System) and No*SQL will be applied in next stage. Future work will study the HDFS framework for supporting enormous data and data mining based DW in the BI system.