Clustering of IT application systems for designing a data warehouse

The successful application of Data Warehouses (DWH) depends on responsible architecture. At this point we should remember Heinz Zemanek, one of the great European pioneers in computing, whose findings overwhelmingly presented the importance of responsible design in IT systems. It was the ideal goal of Bill Inmon to integrate heterogeneous data sources of a company into a Data Warehouse in order to build up a “single point of truth” of data logistics. However, because nowadays companies are building their Federated Data Warehouses without the necessary consideration of the architecture in place, they will never be capable of achieving Inmon’s goal. The question that arises is which IT systems are relevant for Data Warehousing. Yet, the practicable methods required for finding an answer are missing. Applying as a basis the practical experiences of a German power supply plant, this lecture’s proposal is to use the classical clustering methods of IT application systems for a DWH to achieve the data resourcing only of relevant data. Today, with the help of Data Mining, these methods can be implemented faster and more efficiently. Especially Kohonen networks are well-suited for clustering IT application systems and supporting DWH design.