Previous | Table of Contents | Next |
A warehouse administrator needs access to all the necessary information to control and monitor the state of the data warehouse.
To accomplish this, ETL processes need to be scheduled to update information in the data warehouse. Monitoring ETL operations
and journalizing changes to data warehouses must be performed for a variety of data integrity, organizational, and regulatory
reasons. In the event of problems arising, the administrator needs to be able to take appropriate action (such as initiating
a rerun of a set of warehouse processes).
For information held in the data warehouse, the administrator may need to determine its source, derivation, and update history.
This involves identifying transformations that created the information and determining when they last ran. Because the source
of a transformation may itself be another transformation, it may be necessary for the administrator to track backward through
several transformations to identify the original source(s) of the information.
Table 2-2 Application Scenarios
CWM Package |
ETL (Extract, Transform, Load) |
OLAP |
Questionnaire |
Warehouse Administration |
|
Software Deployment | X | X | X | ||
Object-Oriented (UML) | X | ||||
Relational | X | X | X | ||
Record | X | X | |||
Multi-dimensional | X | X | |||
XML | X | X | |||
Transforma-tion | X | X | X | X | |
OLAP | X | X | |||
Data Mining | X | ||||
Information Visualization | X | X | X | ||
Business Nomenclature | X | X | X | X | |
Warehouse Process | X | X | X | ||
Warehouse Operation | X | X | X |