10.4 The Central Metadata Repository | Internet-Enabled Business Intelligence


Team-Fly

	Internet-Enabled Business Intelligence By William A. Giovinazzo
	Table of Contents

	Chapter 10. Common Warehouse Metadata

10.4 The Central Metadata Repository

As we can see, there is a need for accurate and reliable metadata throughout the entire organization. This need extends from the data warehouse architect all the way to the business strategist. Figure 10.4 shows the different groups that require metadata. Each group shown in this diagram has its own unique metadata needs. At first glance, one might not consider end users to be interested in metadata. When we consider our definition of metadatathat it provides the context for the datawe see that users are the ultimate consumers of metadata. Metadata, especially in a BI environment, is critical to the end users. When the user is looking at a particular data element, the metadata provides the meaning. The reverse is true as well: When the user knows what data he or she wants, the metadata tells the user where to find it.

Figure 10.4. Metadata users.

graphics/10fig04.gif

As seen in Figure 10.4, the end user is different from the other users of metadata. The end user consumes metadata but does not have direct input into the metadata repository. The data warehouse architect, system engineers , and administrators all provide input into the repository. The following lists some of the interactions with the metadata repository:

Data Warehouse Architects define the basic structure of the system.
System Engineers enhance the environment that changes the metadata. For example, a review of the data utilization may show that users are consistently asking for certain sets of aggregations. The system engineers may decide to precompile these aggregations. The metadata repository is updated to reflect this change.
System Administrators review performance statistics such as disk and CPU utilization. Administrators modify the environment to improve system performance.
Database Administrators review data utilization statistics and modify the environment to improve performance as well. Database Administrators do such things as build indexes or rebuild fragmented tables.
Data Administrators have perhaps the most interesting interaction with the metadata repository. One might even consider the data administrator the owner of the metadata. The data administrator monitors the quality of the data and detects changes in metadata. It is the responsibility of the data administrator to maintain the quality of not only the data, but of the metadata as well. For example, the data administrator reviews the extraction log. This log detects when the metadata for the operational environment is not synchronized with the data warehouse. The data administrator is responsible for updating the warehouse to allow for proper data loading.

The central metadata repository is critical to the success of not only the data warehouse but of the entire organization. As such, the scope of the repository should extend to the entire enterprise. This enterprise-level view is contained within the Enterprise Data Model (EDM). While the EDM is not strictly part of the data warehouse project, we will address it due to its effects on the warehouse.


Team-Fly

Top