Skip to main content

Different data warehouse models

Short answer

The three data warehouse models are enterprise warehouse, data mart, and virtual warehouse. Enterprise warehouse covers all subjects and is large and costly; data mart is a smaller subject-specific subset, dependent or independent; virtual warehouse is a set of views over operational databases, cheap but without history.

Explain Different Data Warehouse Models.

From the architecture point of view, there are three data warehouse models:

1. Enterprise warehouse

Collects information about all subjects across the whole organisation. It integrates data from many operational systems and external sources, holds detailed and summarised data, and needs a lot of time, cost and planning (GB to TB). Built on mainframes or large servers.

2. Data mart

A subset of corporate data for one group of users, such as marketing. Limited to selected subjects. Faster and cheaper to build (weeks).

  • Independent: loaded directly from operational sources.
  • Dependent: loaded from an enterprise warehouse.

3. Virtual warehouse

A set of views over operational databases. Only some summary views are materialised for efficiency. Easy and cheap to build, but needs extra capacity on the operational servers and gives no historical data.

 Enterprise DW  ->  Data marts (dependent)
 Sources        ->  Data marts (independent)
 Sources        ->  Views (virtual warehouse)

A recommended approach is incremental: build independent marts with a common design, then integrate them into an enterprise warehouse (a mix of top-down and bottom-up).

✓ Answer checked by Claude Opus 5.5 (high effort)

Related questions

All Data Mining chapter 1 questions →