A database warehouse is a collection of databases with information
extracted from other databases. In some cases, end users cannot access corporate
databases for security or legal reasons. In other cases, related information may
be spread across multiple platforms such as mainframes and PC servers.
A database warehouse lets a database manager extract information from such
databases and present it to users as a database whose form may or may not
resemble the original database. A database warehouse can even provide summarized
information.
For example, a corporate sales database contains a list of ordered
products. The customer database contains shipping address information. A product
description database includes product size and weight information. The
product planning group wants the best shipping solution but needs to know the
current requirements. The database manager can use a database warehouse product
to extract the appropriate information from the respective databases, combining
the information into a customized database for the product planning group. This
database may contain only the necessary information, such as the city and
state, plus a summary of units shipped to that area and their average size and
weight. Product planning does not need to know what databases you accessed to
retrieve the information.
Another simple example uses information from one reduced database that
prevents users from accessing the private information. For example, an employee
database could be the basis for an employee telephone directory. Anyone could
access this part of the database, but the employee database with private
information, such as salary and benefits, would remain secure.
A database warehouse typically includes a meta-directory, or information
catalog. A meta-directory describes what data the database warehouse maintains,
including the data type (integer, string, etc.) or column in a database. The
meta-directory lets database warehouse users determine what databases to look
at, based on their content.
The meta-directory presentation often lets users select both the source of
information and an analysis and presentation method. Users often generate
reports repeatedly, supplying new data or different selections. Alternatively,
you can create custom applications to use the meta-directory to locate
information, and the same application can then access the selected database.
A meta-directory typically consists of a database that the database
warehouse extraction support maintains. You can adjust the presentation
of the meta-directory to display information in terms that users can appreciate,
including annotations such as the original location of extracted database
information. Some database warehouse products or database servers use the Web
server to present this information.
Database warehouses operate on a simple principle: Take information from
one set of databases and put it in another. This process is an extraction. As
always, the devil is in the details. The destination databases do not exist in
realtime. You must define and schedule each extraction, and you can extract
information frequently (e.g., hourly) or infrequently (e.g., once a month).
Database managers often organize database information for a specific
application or set of applications. This practice may make the database
unsuitable for human consumption or analysis. Database warehouse software lets a
database manager become an information manager while letting users use the
warehouse information directly.
End of Article
