Data Vault Automation in a truly business driven approach

Michael Müller, MID

In Data Vault the actual data integration is done on business keys. Business Keys identify business objects across different departments of a company. Integrating data based on business keys delivers a solid business object model of the company (or a part of the company).
When there are no business keys, the building of the business key, the integration of the data must be done in the Data Warehouse. With Data Vault this often leads to a ‘source driven’ approach. A direct mapping of the source system data model.
Especially with a data warehouse automation approach this can be done very fast. Then the integration needs be done in the business vault and the resulting data vault model is huge and barely understandable to the business users. Communicating the problem of the underlying data and how it needs to be integrated into a common business view is very hard. Every model that is shown is just too complicated with too many entities. It is a very technical data model.
With a business-driven approach, the starting point will be a business object model. This is a good and easy model for Business and IT to build a common understanding of the underlying data. This business object model can then be used to generate a data vault model. Now the source must be mapped to this business-driven data vault and still can be automated. And by starting with a business object model there is no elephant in the room. Right from the beginning the communication between Business and IT is on point about integrating the data.
We will cover the following topics:

  • Source vs. business driven
  • Business object model and business keys
  • The value of a business object model
  • Differences in the automation process
  • Data Integration: common challenges and solution patterns

Michael Müller models and designs Data Warehouses now for 15 years. He works currently as principal consultant for MID and usually works as link between user and IT. Starting from requirement and stopping with the design of the warehouse. As part of a huge longtime data warehouse project he knows about the impacts of changes to a warehouse firsthand. He is very passionate about using metadata to speed up the development process and automation in general.

Comments are closed.