Abstract

Enterprise or group data warehouses are often introduced
in complex multi-national organizations in order to foster
harmonization, integrate heterogeneous source systems and
hide the heterogeneity from analytical systems. Industry
reference data warehouse logical data models such as Teradata's
FS-LDM or IBM's BDWare promoted as accelerators
for the development of such large data warehouses. However,
this paper shows that logical data models alone are
not sucient to ensure reusability in those environments.
In order to provide a solid basis for standardization, the
logical data model needs to be accompanied by a semantic
business information model used as an anchor point for
the mappings and for communication with business users.
Such a model allows a model-driven approach for speci cation
of the data transformations, which usually accounts for
at least half the total e ort of large data warehouse projects.
The paper presents an approach building upon the Teradata
Business Data Element (BDE) concept giving practical examples
from project experience in the nancial services industry.
A research prototype is presented, utilizing Semantic
Web technologies such as the Web Ontology Language
(OWL) to facilitate the traceability of data requirements,
business terms and physical data elements in the di erent
layers of a complex data warehouse architecture.

Download

Share

COinS