Want to break down the barriers? Tired of "taking sides" when you don't have to? In this blog I explore a modeling technique called the Data Vault (no it doesn't have to do with security or locking your data away). This technique sits squarely between Inmon 3rd normal form warehouse and Kimball Star Schema design as a warehouse.
This modeling technique is comprised of the best-of-breed from both designs and is built to overcome limitations of the adaptations made to each data modeling architecture; specifically with regards to data warehousing.
What is a Data Vault?
Definition: The Data Vault is a detail oriented, historical tracking and uniquely linked set of normalized tables that support one or more functional areas of business. It is a hybrid approach encompassing the best of breed between 3rd normal form (3NF) and star schema. The design is flexible, scalable, and consistent. It is a data model that is architected specifically to meet the needs of todayâ€™s enterprise data warehouses.
What is it's real name?
Common Foundational Warehouse Architecture
What does it do for me?
* Saves time and money in build out of your data warehouse or enterprise integration initiative.
* Provides a consistent, repeatable architecture that can scale to your enterprise needs.
* Defines easy to use standards for all to follow
* Reduces complexity of the integration effort
* Saves storage space
* Increases visibility into both: well-oiled and broken business processes
* Demonstrates a strong basis for Data Visualization and Data Mining
What are some of the benefits?
* Rapid prototyping and build out of data marts and reporting solutions.
* Genuine audit trail pictures produced of the enterprise vision of data (even if none exist on the source systems)
* Data is modeled by type, and rate of change - allowing both batch and real-time to be "added" to the warehouse at the same time.
* Contains a high degree of data attribution
* Scales to petabyte levels if necessary
* Can handle near real time data arrival at a fraction of a second.
Yes, but are there any customers using it?
Sure - just check the web site for more information.
What are some of the success stories?
* Large manufacturing company saved millions when finding and fixing a billing error that had been occurring for the past 15 years.
* Large financial company integrated M&A 3 companies in 3 months flat
* Large banking company adds "branches" to their warehouse quickly at a low cost.
* Government operation decreases "time to build data marts" to 1 hour (from requirements to inception).
What makes the Data Vault so successful?
It's ability to be modeled AT THE BUSINESS LEVEL. The Data Vault is designed to mimic the business keys - the most important data element in business. Once the keys are established it models the relationships across those keys; which flush out both process relationships and undocumented business operational relationships. Finally the attribute data or descriptive data is added to the mix and split by Type of Data and Rate Of Change.
From a business perspective, the logical model is tied tightly with the physical model and architecture - therefore it is easy to change the model as the business changes. It is also based on a statement of fact. The business keys are in use and were in use at a specific point in time. This model is built to capture those ideas.
Come hear more about this technique at TDWI, Sunday October 30th in Orlando. Or contact me directly: daniel.Linstedt@myersHolum.com
Posted September 27, 2005 5:26 AM
Permalink | 3 Comments |