Databases Reference
In-Depth Information
Why is MDM important? In the traditional world of data and information management, we used to
build data and applications in silos across an enterprise. The addition of new systems and applications
resulted in not only data volumes and transactions, but also created redundant copies of data, and in
many cases the same data structure contained disparate values. The end-state architecture resulted in
systems that did not interface and integrate with each other. The complexity of processing disparate
data into a common reference architecture required hours of manual effort and did not provide a clean
and auditable result set. Each system could give you a fractured insight into what was happening in
the enterprise, but you could not create a clear and concise view of data as a centralized system.
This is where the need for a centralized MDM system begins. With a centralized system, the enter-
prise can create, manage, and share information among systems and applications seamlessly. The efforts
to manage and maintain such a system are very simple and flexible compared to a decentralized plat-
form. This approach can save an enterprise time and opportunity costs, while ensuring data quality and
consistency. MDM is driven to handle each subject area as its own system within the enterprise. The
underlying architecture of the system allows multiple source systems to be integrated and each system
can alter the attribute values for the subject area. The final approval of the most accurate value is deter-
mined by a data steward and a data governance team, after which the business rules are executed to
process the data modifications. The results are then shared back with the source systems, applications,
and downstream analytical applications and databases, and called the “gold copy” of the data definition.
MDM is not about technology. The critical success factor of this initiative is the subject matter
experts in data within the business teams, who can understand and define the processing rules and
complex decision-making process regarding the content and accuracy of the data. MDM is not imple-
menting a technology; as the role of any technology platform in this process, it is that of a facilitator
and an enabler.
MDM is about defining business processes and rules for managing common data within dispa-
rate systems in an enterprise. In implementing these processes, the data governance and stewardship
teams collectively determine the policies, validation, and data-quality rules, as well as service level
agreements for creating and managing master data in the enterprise. These include:
●
A standardized definition of data common to all the systems and applications.
●
A standardized definition of metadata.
●
A standardized definition of processes and rules for managing data.
●
A standardized process to escalate, prioritize, and resolve data processing issues.
●
A standardized process for acquiring, consolidating, quality processing, aggregating, persisting,
and distributing data across an enterprise.
●
A standardized interface management process for data exchange across the enterprise internally
and externally.
●
A standardized data security process.
●
Ensuring consistency and control in the ongoing maintenance and application use of this
information.
Master data sets are used across transactional databases, operational databases, web applications
and databases, data warehouses, datamarts, and analytical databases. There are several techniques to
implement a master data repository and, depending on the nature of the applications and databases
involved, enterprises choose their custom implementation process. MDM transcends all architecture
schools of data warehousing and is not exclusively confined to dimensional models alone. The two
