Analysing multi-dimensional data across autonomous data warehouses

Abstract

Business cooperations frequently require to analyse data across enterprises, where there is no central authority to combine and manage cross-enterprise data. Thus, rather than integrating independent data warehouses into a Distributed Data Warehouse (DDWH) for cross-enterprise analyses, this paper introduces a multi data warehouse OLAP language for integrating, combining, and analysing data from several, independent data warehouses (DWHs). The approach may be best compared to multi-database query languages for database integration. The key difference to these prior works is that they do not consider the multi-dimensional organisation of data warehouses. The major problems addressed and solutions provided are: (1) a classification of DWH schema and instance heterogeneities at the fact and dimension level, (2) a methodology to combine independent data cubes taking into account the special characteristics of conceptual DWH schemata, i.e., OLAP dimension hierarchies and facts, and (3) a novel query language for bridging these heterogeneities in cross-DWH OLAP queries. © Springer-Verlag Berlin Heidelberg 2006.