Repository Support for Data Warehouse Evolution

Data warehouses are complex systems consisting of many components which store highlyaggregated data for decision support. Due to the role of the data warehouses in the daily business work of an enterprise, the requirements for the design and the implementation are dynamic and subjective. Therefore, data warehouse design is a continuous process which has to reflect the changing environment of a data warehouse, i.e. the data warehouse must evolve in reaction to the enterprise’s evolution. Based on existing meta models for the architecture and quality of a data warehouse, we propose in this paper a data warehouse process model to capture the dynamics of a data warehouse. The evolution of a data warehouse is represented as a special process and the evolution operators are linked to the corresponding architecture components and quality factors they affect. We show the application of our model on schema evolution in data warehouses and its consequences on data warehouse views. The models have been implemented in the metadata repository ConceptBase which can be used to analyze the result of evolution operations and to monitor the quality of a data warehouse.