|
 |
Approaches for Data Mart Implementations
 by Arshad Kahn
IT Professional, Independent Consultant, and Author
June 27, 2007
A data mart aims to meet the needs of departments and smaller groups, rather than the complete enterprise. The design principles and objectives of a data mart and a data warehouse are the same. While both are decision support tools and have the same basic characteristics, such as being subject-oriented and integrated, a data mart is far smaller in size and scope.
A data mart has far less users and the size of its database, typically, is only a few gigabytes. This is relatively small compared to the typical data warehouses (storing hundreds of gigabytes) or the larger data warehouses (containing terabytes of data). A data mart requires simpler hardware and supporting technical infrastructure. Hence, it can be implemented by staff with less experience and technical skills.
Implementation Approaches
There are two primary approaches that are commonly used to build data marts, and the one you should choose depends primarily on available resources and time to implement.
- Top-down: Build an enterprise data warehouse (EDW) and then construct dependent data marts, which are its highly summarized subsets
- Bottom-up: Build independent data marts, whose foundation is the enterprise data model, which can then be used to construct an EDW
Next...The Top-Down Approach
Page 1 2 3 4
Related Articles
|
|
|