Recursive decomposition as a method for integrating heterogeneous data sources

Abstract

In this article, proposes a model of data integration in which should be supported a unified view of heterogeneous data sources, management of integrity constraints, managing the execution of operations of data manipulation and query, reconciliation of data from different sources, the ability to extend and customize the new data sources. The proposed approach to integration is based on the recursive decomposition of data sources where each source data is divided sequentially into atomic data items, wherein for each level of the recursive nesting data and their descriptions are represented uniformly. Such model enables the integration of different data sources at any level by setting arbitrary links between elements of the scheme, integrity constraints and allowed operations.