Databases Reference
In-Depth Information
the first chapter of this topic, we saw how each datum
duplicated between the systems is subject to governance by
the MDM solution; we have also explained that this concerns
the exchange of data between companies. The MDM system
ensures the traceability of these exchanges and verifies the
integrity of the information before making it available to
other systems. To achieve this, the MDM system must
automatically enforce referential integrity constraints
through data modeling. If the data model is impoverished, as
in this first version of the address model, the MDM system
has no jurisdiction. So then, bespoke and hard-coded
software must, at EAI-ESB level, be developed in order to
analyze the data flows, and strictly enforce the rules for
validation.
The more impoverished the data model is, the more
bespoke software development is necessary to restore the
integrity of the data. This design is expensive and locks
business knowledge in hard-coded software. The objective of
semantic modeling is to restore power to the data model in
order to reduce the need for this type of design.
With an address model of this type, it is impossible to
ensure a real traceability of data flows exchanged between
systems. By memorizing the addresses with the help of this
model, we can obtain a technical audit trail of data flows
only, since the relationships between the data are not fully
expressed. It is thus complex to execute data queries of the
type “get a list of all the addresses exchanged with a
company whose town is in a given country”. The solution
necessitates a lookup of the totality of the addresses because
there is no relationship between countries (“Country”) and
cities (“CityName”).
To complete the critical analysis of this first version of the
model, we must also take note of the absence of names of the
associations. In this data model, quite short, it is easy to
guess the meaning of the associations; this is no longer
