Database Reference
In-Depth Information
this global framework can make such integration dicult and costly in the
long term.
There is no consensus on the phases that should be followed for data
warehouse design. Some authors consider that the traditional phases of
developing operational databases described in Chap. 2 , that is, requirements
specification, conceptual design, logical design, and physical design, can also
be used in developing data warehouses. Other authors ignore some of these
phases, especially the conceptual design phase. Several approaches for data
warehouse design have been proposed based on whether the analysis goals,
the source systems, or a combination of these are used as the driving force. We
next present these approaches, which we study in detail in the next sections.
The analysis-driven approach requires the identification of key users
that can provide useful input about the organizational goals. Users play
a fundamental role during requirements analysis and must be actively
involved in the process of discovering relevant facts and dimensions. Users
from different levels of the organization must be selected. Then, various
techniques, such as interviews or facilitated sessions, are used to specify
the information requirements. Consequently, the specification obtained will
include the requirements of users at all organizational levels, aligned with the
overall business goals. This is also called goal-driven approach.
In the source-driven approach , the data warehouse schema is obtained
by analyzing the underlying source systems. Some of the proposed techniques
require conceptual representations of the operational source systems, most of
them based on the entity-relationship model, which we studied in Chap. 2 .
Other techniques use a relational schema to represent the source systems.
These schemas should be normalized to facilitate the extraction of facts,
measures, dimensions, and hierarchies. In general, the participation of users is
only required to confirm the correctness of the derived structures or to identify
some facts and measures as a starting point for the design of multidimensional
schemas. After creating an initial schema, users can specify their information
requirements. This is also called data-driven or supply-driven approach.
The analysis/source-driven approach is a combination of the analysis-
and source-driven approaches, which takes into account what are the analysis
needs from the users and what the source systems can provide. In an ideal
situation, these two components should match, that is, all information that
the users require for analysis purposes should be supplied by the data included
in the source systems. This approach is also called top-down/bottom-up
analysis.
These approaches, originally proposed for the requirements specification
phase, are adapted to the other data warehouse design phases in the method
that we explain in the next section.
Search WWH ::




Custom Search