Page 56 -

A Company and its Data 17

Finally, a scattered IT system, under the influence of data
duplications, is an unrecognizable system. Knowledge of the
data and their relationships is diluted at several points in
such an Information System, and only the data integration
layer is at hand in order to try and achieve some unified data
validation. This is misleading, however, because this
integration is not based on a real data model in the
traditional sense. It remains technical (pivot format),
without transparency, vis-à-vis the business users who
should be more involved in the management of the necessary
validation rules.

In order to rectify this situation, it is important to have a
data repository that is used by the integration layer. This
data repository embodies a data model that jointly expresses
the content of business objects used by data exchanges
between systems, as well as the totality of their validation
rules and their relationships with one another (referential
integrity constraints). This is a rich data model.

This data repository is accompanied by governance
functions that enable data to be queried, using business
friendly terms, to access audit trails and data traceability.
These functions enable successive versions of objects to be
stored in order to follow data exchanges over time, compare
versions of the same object, or to restart a process of data
synchronization at a known version, etc.

These functions are not just at the disposal of IT experts,
as this is not an additional technical supervision console. On
the contrary, it is a tool designed for business users (see
Figure 1.2).

This data repository is an MDM system used for the range
of data duplicated between systems. The data in question