Page 71 -

32 Enterprise Data Governance

for B2B third parties (individuals, citizens) does not allow
this all of the time. In that case, it is best to couple a
transactional hub with MDM.

Indeed, there are a number of reference and master data
that define customers. For example, classifications,
hierarchies, incorporation of customers within geographical
sectors or socio-economic categories, customer risk levels, etc.
All this data must be governed even before transactions are
executed in the hub. On the other hand, certain transactions
carried out at the CDI level could cause rejection of
information that is in one’s best interest to submit to the
MDM system in order to benefit from its data governance
functions. For instance, if the CDI detects that a client is not
connected to a mandatory classification, then it can send a
part of the file to the MDM system so that a data steward
can complete it.

Model-driven MDM Data Injection CDI (OLTP system)
Governance time To data governance Run time
Rich data model Rigid data model
over OLTP RDBMS in OLTP RDBMS

RDBMS Tables RDBMS Tables
Business objects, lifecycles, Consolidation and
data enrichment
validation rules, permissions

All data governance functions
Data Steward are automatically driven by Transactional IT System
the rich data model
Usual users of the
IT system

Figure 1.5. Collaboration between CDI and MDM

A collaboration therefore appears between CDI and MDM
(see Figure 1.5):