| Data warehouse is a good solution to manage huge information resources. It can integrate a wide range of business data through data consolidation technology ETL (Extract, Transform, Load) and provide a basis for effective usage in future. However after building the data warehouse platform, some business core master data are still very difficult to be consolidated as because of its multi- data sources, diverse and complex business logics and etc. Therefore further data analysis, data representation, data mining and other applications will cause problems accordingly.In order to solve above problems, and also for maintaining greater data consistency through various systems, master data management (MDM) comes into existence. MDM is a combination of tool, process, technology and management. It can help enterprises create a more effectively single, accurate and authoritative source of information.This article describes the definition, characteristic, relationship of master data, MDM and data warehouse. It introduces the system architecture for MDM and related detail design for its application layer, master data repository and master data integration service.Through introduction for one project practice of MDM applied in data warehouse, including its project background, master data analysis, system architecture, detail design and project conclusion, it gets the point that using an effective combination for MDM, data warehouse will be able to access shared, complete and more accurate data and enterprises in the field of business intelligence will also take more obvious advantages. In addition, different from data warehouse one-way integration (multiple systems' data into a single storage). MDM is able to build a necessary two way integration, which is able to synchronize those maintained master data back to its related source systems. This approach will achieve the objective for master data being consistency, integrity and controllability in the whole enterprise wide. |
PDF Full Text Request