Databases Reference
In-Depth Information
7.1.4.1.
Is the mutualization of data modeling dangerous?
Let us try to understand the risks surrounding the
mutualization of the model using an example, that of the
“Product” data repository. When a product is configured it
must be attached to the factories which fulfill the criteria
that fit its configuration.
In order to decide which factories to attach, the MDM
system presents the user with a set of data concerning its
workshops, its production capacity and other means of
supply. The data model for the configuration of products is
thus linked to another data model, that of the description of
the factories.
The business object “Product” is under the responsibility of
the “Product management” department and the business
object “Factory” is under the responsibility of another
business unit, that of the “Manufacturing department”. The
consequences of the lack of reuse of the “Factory” data model
between these two business units will now be described, as
will the consequences of its mutualization.
The Factory data model without mutualization
Without mutualization, the product data repository
includes its own modeling for factories. This modeling is rich
because it contains all the characteristics necessary for the
selection of factories according to the configuration of the
products. Data values for this model must be loaded by the
factory data repository which is under the responsibility of
the manufacturing department. This second data repository,
due to the absence of mutualization in modeling, has its own
data model. The IT function has then to put in place a data
synchronization mechanism between the two models
describing factories, the one included in the MDM “product”
and that of the MDM “factory”. IT has to synchronize the two
separate databases even though the description of the
factories could be common to both. The technical workload is
