Databases Reference
In-Depth Information
MDM system can be insufficient to fix data integrity
problems. Such an MDM system would therefore be
completed by “hard coded” software developments, freezing
the data validation rules that depend on the states.
8.1.1.3.
Modeling in a comprehensible, stable and sustainable
manner
The data models are structured progressively during the
production of the MDM system. We will see that this
approach, in stages, is based on a structure around three
architecture levels that we call “business object domains”,
“data categories“ and “business objects”. In the absence of
this distinction in the architecture, the sharing of data
between actors and systems in the company becomes harder.
Equally, the ability of the models to be extended or
maintained is then harder to guarantee.
This structure is robust, negotiated and shared across the
whole Information System, and even beyond if the data is
being exchanged with other organizations. The detail of the
data, inside each business object, can evolve depending on
what is needed, but the structure forms a stable and
sustainable asset. This architectural building gathers
together disciplines varied in the methods, organization and
knowledge of business. It is a multi-disciplinary task,
unavoidable as soon as the aim is an implementation of an
MDM system across the whole of the information system. We
must ensure that its construction is achieved by best
practices and that putting it in place occurs in a progressive
manner.
8.1.1.4.
Modeling setting aside the technical constraints
The MDM system provides the other systems with the
reference and master data, either by replication of the data,
or by a service mode access (see Chapter 12). Some data is
updated by systems other than the MDM system. It is then
necessary to envisage a synchronization process between the
