The architecture wars have raged for years about when and why different types of data warehouse solutions should get built. Some advocate an interventionist approach in which enterprise data warehouses are developed that are fundamental to the enterprise, while others recommend a softer touch with departmental business intelligence solutions and virtual integration of data sets.
All of the approaches have one thing in common: In some way, they duplicate at least the business rules and almost always the data associated with individual business application instances across the enterprise.
At first glance, this may appear counterintuitive. If the objective is to have a consistent view of business data, then duplicating the content introduces multiple opportunities for error. In addition, many senior technology executives are concerned about the cost of storage with ongoing duplication of content; much of this concern is due to their experiences as middle managers in the 1980s and 1990s when such storage was excessively expensive. The cost of storage should seldom be an issue today.
The data warehouse seems to have come into being without any single inventor, although system architects at Digital Equipment Corporation (DEC) and IBM both lay claim through the 1980s to early development of the concept. IBM in particular invested in the concept of a so-called "information warehouse" providing an integrated view of the enterprise. ...