A
good database design is the cornerstone of data management and data
architecture, supporting:
The
ideal starting point for database design is data
modelling: The essential data structures of a business must be clearly
understood and shared. The data model can then be implemented as one of the many
commonly used relational databases. State of the art modelling
tools facilitate the transformation of a logical data model into a
database-specific physical data model. Performance and storage considerations
are key factors of this transformation. The resulting physical data model is
then used to generate the database. Good modelling tools support both
incremental changes and round-trip engineering, ensuring that data model and
database cannot get out-of-synch.
Data
models often used to be design-time documentation, which became soon out of
date, and were therefore not a reliable source of data structure information.
This has changed with the advent of advanced modelling tools. Synchronized data
models are now both design- and run-time documentation, guaranteed up-to-date,
and augmented with descriptions for classes, attributes and associations. Such a
data model is an ideal candidate for web publishing, on the IntraNet for
example, to serve as a graphical, searchable data dictionary for business users
and IT specialists alike.
The
change management capabilities of the modelling
tools keep track of the changes and versions over time. This for example can
greatly assist in determining the delta between two versions, or the delta
between a test environment and a production environment.
With
over 20 years experience in data modelling, database design for various
mainframe to mid-range relational databases and the most advanced modelling
tools, I am well positioned to let you reap the benefits of model-driven data
management and database design.
|