Data Vault Ensemble Modeling

Layering Business Logic on your Data Vault

Roelant Vos, Allianz Worldwide Partners

The Data Vault methodology provides elegant concepts to develop your Data Warehouse – the various required Data Warehouse mechanics are organised in a way that allows for a flexible solution.

By and large, it is fair to say the pattern archetypes such as Hubs, Links and Satellites are sufficiently understood by the broader community.

Indeed, there have been many great examples of how metadata (model) driven approaches simplify delivery. These approaches, both open-source and commercial, leverage the templates and insert the required information from repositories or domain-specific languages.

When this foundation is in place and a baseline for rapid development, deterministic refactoring and deployment has been established, the dynamic shifts to aligning the data with the business’ expectations.

This session discusses the concepts that are available to apply business logic on a Data Vault (and where to do this), as well as examples how this can be implemented.

Contents:

  • Separation of concerns approach for Data Vault; what is already in place ‘out of the box’?
  • Different approaches towards capturing transformation (business logic) metadata
  • Incorporating business logic in ETL generation
  • Abstracting levels of design; how far can we go?
  • Tips to keep your solution manageable
  • The role of ‘time’ in delivering outputs the business can work with

During the day I work as General Manager – Enterprise Data Management (which roughly covers Data Governance, BI and Analytics) and at night I try to improve the status quo in the data community through open source development and framework collaboration. With this I hope to contribute to the delivery of enterprise wide data management that is fast, flexible, future proof and easy to manage.

To me, working with data is endlessly varied and finding more ways to further simplify (‘automate’) data management continues to be a source of inspiration.

Because of its automation potential and clearly defined patterns I have been a fan of Data Vault and similar hybrid approaches for Data Warehouse design for almost 15 years. I have written many articles around this theme on www.roelantvos.com/blog and also delivery my own Data Vault implementation and automation classroom training.

I am a practical person with many interests and a strong technical background and would like to get (and stay) in touch with like-minded professionals. Please reach out on roelant@ravos.nl or participate on www.roelantvos.com/blog to see where we can continue to improve our data management solutions.