We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.


The Good Data Mart

Originally published April 7, 2011

The Pandora’s box that is an independent data mart is well documented. All sorts of ills beset the organization that builds and tries to support the independent data mart. There is the lack of integrity of data. There is the interface that consists of a zillion programs that have to be written, maintained, and executed on a nightly basis. There is the redundancy of data that exists from one independent data mart to the next. And this is just the short list of what befalls an organization that tries to base its information architecture on an independent data mart foundation.

But just because there are many architectural pitfalls with an independent data mart, does that mean that we shouldn’t be building a data mart at all? No it does not. When the data mart is a dependent data mart, there are many benefits that will ensue.

A dependent data mart is one where all the data in the data mart comes from a data warehouse. In the data warehouse, there is a foundation of granular, integrated historical data. That foundation data is used to feed the different data marts. The data warehouse granular data is aggregated, summarized, and restructured as it passes into the data mart. The end result is a data mart that is customized to meet the needs of the end user.

So what are some of the benefits of a dependent data mart? It turns out that there are many benefits:

  • By building a dependent data mart, a burden of processing is lifted from the data warehouse. Stated differently, when performance becomes an issue in the data warehouse, building one or two dependent data marts is one of the best things the analyst/designer can do. By doing processing outside the data warehouse, performance inside the data warehouse suddenly becomes a lot better.

  • By building a dependent data mart, the designer analyst can tailor the dependent data mart to the specific needs of the end user analyst. The data in the dependent data mart can be reshaped and restructured into a form that very specifically meets the needs of the end user.

  • By building a dependent data mart, the linkage/lineage from one data element to the next – the heritage of data – can be tracked. When an end user wants to see where a data element has come from, the end user analyst can trace the lineage of data where there are dependent data marts.

  • By building a dependent data mart and by placing the dependent data mart in a separate processing facility, the economics of analytical processing can be greatly reduced. Nothing is more expensive than the cycles of processing that occur on a large centralized processor. By building a dependent data mart and by moving the dependent data mart to an external processor, the costs of processing can be dramatically reduced. No longer does all processing have to occur on an expensive centralized processor.

  • By building a dependent data mart, key performance indicators (KPIs) can be created and tracked over a long period of time. The dependent data mart is the ideal place for the building and tracking of KPIs.

  • By building multiple dependent data marts, there is always a single point of reconciliation for all dependent data marts. When the analysis of one dependent data mart conflicts with the analysis created by another dependent data mart, there is always the granular data in the data warehouse that serves as a single point of reference for reconciliation of the differences between the analysis created by different dependent data marts.
And the list goes on. There are MANY advantages to the creation and operation of the dependent data mart. But perhaps the single largest advantage to the dependent data mart has nothing to do with technology or architecture at all. Perhaps the single largest advantage to the dependent data mart is the fact that different departments within an organization can own their own data. Departmental units of an organization love to own something of their own and love the control that comes with ownership.

When a department owns its own dependent data mart, that department can shape the KPIs any way they want. They can do any kind of processing that they want. They can add other analysis any way they want. Owning the dependent data mart implies that there is freedom to do whatever the end user department deems necessary and useful.

So there are then some very powerful and compelling reasons why organizations choose to build and support multiple dependent data marts.

  • Bill InmonBill Inmon

    Bill is universally recognized as the father of the data warehouse. He has more than 36 years of database technology management experience and data warehouse design expertise. He has published more than 40 books and 1,000 articles on data warehousing and data management, and his books have been translated into nine languages. He is known globally for his data warehouse development seminars and has been a keynote speaker for many major computing associations.

    Editor's Note: More articles, resources and events are available in Bill's BeyeNETWORK Expert Channel. Be sure to visit today!

Recent articles by Bill Inmon

 

Comments

Want to post a comment? Login or become a member today!

Be the first to comment!