Summary

Data Vault modeling is a mix between normalizing data and dimensional modeling. It is designed to provide a flexible way to store detailed, historical data. By using Hubs, Links, and Satellites, you create a database in which you never need to alter an existing table. All changes can be handled by adding new tables. That makes a Data Vault more stable over time than other database designs.

Data Vault is also very standardized. This enables tools to automatically generate Data Vault structures from existing normalized databases. On top of that, the ETL process that loads the tables can also be generated for the most part.

As a third benefit, Data Vault is scalable. With the possibility to load all tables in parallel, it can leverage powerful hardware and load vast amounts of data in short periods of time.

Using the business vault functionalities improves the usability of data. However, using Data Marts behind the Data Vault is still a best practice.

A Data Vault...