Incremental Integration of Data Warehouses: The Hetero-Homogeneous Approach

As organizations grow, decision support systems face a surging demand for information; information that must be integrated into the corporate data warehouse. Incremental integration allows for a gradual evolution of data warehouse schemas.

Heterogeneous schemas complicate the integration task. Traditional modeling approaches require heterogeneities in the data warehouse schema to be eliminated, or they lack expressiveness when capturing heterogeneous information. When dwelling on homogeneous schemas, however, valuable information is lost. In contrast, the hetero-homogeneous modeling approach preserves the clarity of homogeneous schemas while offering the flexibility that is needed to model large companies with diverse sub-entities.

In this paper, we demonstrate the application of well-accepted object-oriented design principles to data warehousing. We examine the incremental integration of data warehouses, and we sketch the fundamentals of our prototype implementation. Following the hetero-homogeneous approach, we preserve schema heterogeneities as much as possible while avoiding common pitfalls of heterogeneous data warehouse modeling.