Databases Reference
In-Depth Information
5.2. Static MDM
Static MDM is most widely used in companies today. It
can also be the most dangerous because it is not always
enough to ensure full data integrity. It is not enough just to
put in place MDM processes to automatically increase the
quality of data. If the data repository relies on a model which
does not adequately include data validation rules, it can lead
to increased data quality issues. For example, in a financial
data repository, if a business rule forbids the creation of
more than three levels of financial classification, the MDM
system must formally include this rule. If this rule becomes
more complex, because the levels of classification change
depending on the status of the financial operations, a more
sophisticated data model, including validation rules
depending on status is required.
In the absence of rich modeling, the MDM system only
handles the static view of the data, which necessitates a
hard-coding software development to implement data
validation rules.
In this situation, the MDM system and its ad hoc heavy
and hard-coded routines located outside the data model,
becomes an additional application silo needing complex
software maintenance. This MDM system becomes weighty
and difficult to maintain.
A static MDM system, as opposed to a virtual MDM
system, contains the reference and master data of business
objects. It does not limit itself to managing the identifiers
and associations between objects, nor a frozen header record
at the moment of creation. This data model allows the
storing and update of all reference and master data without
restrictions. It is a true data warehouse. This MDM system
allows updates, which means it must ensure a complete
control of the data validation process. Business rules are
required to check the integrity of data, both in pre-condition
