Databases Reference
In-Depth Information
A.3. Semantic modeling and data quality
The modeling that we have described considerably
increases the data quality of addresses, as soon as the data
has been input into the system. Due to the fact that each
address is attached to an abstract
Object
, the identification
of which is unified across the whole Information System
without duplication, the risks of creating a duplicate address
disappear. Unlike quality tools that clean up existing
databases, semantic modeling establishes a new repository,
the quality of which is fully enforced as soon as the data is
initialized and updated.
The usual quality tools are still useful to clean up certain
residual imperfections that can occur at the time of data
input, but more simply, by acting on the MDM repository
rather than legacy databases.
A.4. Performance
The response times to read an address in the MDM
repository depend on the tool used. The indexing and
optimization mechanisms enable creation of indexed views
which collect, virtually, several semantic classes in the same
logical table. The cohabitation between Model-driven MDM
and classic OLTP databases is common. From then on, the
data governance applied to address management is carried
out in the MDM repository to then be injected into the OLTP
databases that use them. In all cases, the MDM system
remains the reference (one version of the truth) that
influences all other systems (see also Chapter 12).
A.5. Lifecycle of the
Address
business object
The
Address
semantic class contains a set of extended
business operations stemming from its state machine that
defined its lifecycle. We have already described this dynamic
