Page 89 - DCAP208_Management Support Systems

Management Support Systems

Notes
Figure 6.4: Three-layer Architecture for a Data Warehouse System

Source: http://www.mhprofessional.com/downloads/products/0071610391/0071610391_chap01.pdf
Finally, let’s consider a supplementary architectural approach, which provides a comprehensive
picture. This approach can be described as a hybrid solution between the single-layer architecture
and the two/three-layer architecture. This approach assumes that although a data warehouse is
available, it is unable to solve all the queries formulated. This means that users may be interested
in directly accessing source data from aggregate data (drill-through). To reach this goal, some
queries have to be rewritten on the basis of source data (or reconciled data if it is available). This
type of architecture is implemented in a prototype by Cui and Widom, 2000, and it needs to be
able to go dynamically back to the source data required for queries to be solved (lineage).

6.3.4 An Additional Architecture Classification

The scientific literature often distinguishes five types of architecture for data warehouse systems,
in which the same basic layers mentioned in the preceding paragraphs are combined in different
ways.
In independent data marts architecture, different data marts are separately designed and built in
a non-integrated fashion (Figure 6.5). This architecture can be initially adopted in the absence of
a strong sponsorship toward an enterprise-wide warehousing project, or when the organizational
divisions that make up the company are loosely coupled. However, it tends to be soon replaced
by other architectures that better achieve data integration and cross-reporting.

82 LOVELY PROFESSIONAL UNIVERSITY