Databases Reference
In-Depth Information
unnecessary complexity, i.e. without direct relation to the
data exchange needs.
This weakness is inevitable if the teams in charge of the
ESB do not have the means or the competency to lead data
modeling, across the whole of the information system, to
fruition. This is almost always the case; the ESB is perceived
as a technical solution that leaves no room to budget for data
modeling. With the MDM approach, the situation changes
radically. Indeed, semantic modeling becomes unavoidable.
The effort dedicated to data modeling, across the whole
information system, is justified by the MDM approach. Even
though it is only concerned with reference and master data,
it requires, beforehand, the implementation of a common
enterprise data architecture that embraces reference/master
data and transactional data at the same time (see Chapter
8).
From then on, since a rich data model is available, it
becomes evident that it is necessary to reuse it instead of
usual pivot formats which have inferior data quality. This
has the following advantages:
− since the data model is based on a common enterprise
data architecture foundation, its stability over time is
reinforced. The business objects domains, data categories
and business objects are stable and shared across the all the
information systems (see Chapter 8). The degradation of the
pivot formats, caused by a progressive stratification of data
structures, no longer exists;
− we have seen that the maturity level of a “semantic
MDM” system takes into consideration the dynamic
modeling of data, in particular the lifecycles of business
objects and data validation rules. The classic hard-coded
software development of the data validation rules outside the
pivot formats is reduced. Consequently, the rich data model
