The Microsoft® Data Warehouse Toolkit: With SQL Server™ 2005 and the Microsoft® Business Intelligence Toolset

13.2. Metadata Standards

While projects to build an enterprise repository are often less than successful, the effort continues because there are at least four good reasons to have a standard, shared repository for metadata.

• When tools can exchange metadata, you can reuse existing metadata to help define each new step in the implementation process. Column names and descriptions captured in the data model design step can be used to build the relational tables, reused to populate the OLAP engine, and then used again to help populate the front-end tool's metadata layer, for example.

• If the metadata is in a standard form, your investment in defining objects and processes is protected—you are not locked in to a particular tool. For example, Reporting Services Report Definition Language (RDL) is a language standard that describes a report. If all your reports are stored in RDL, you could potentially switch to a new front-end tool and still be able to use your existing report library. (This benefit is not particularly popular with tool vendors.)

• A central repository gives you a single, common understanding of the contents and structure of the data warehouse—it is the best documentation you could have. And, to the extent that this shared repository holds the official, active metadata for each tool, you know it is the current version and not a copy that may be out of date.

• An integrated metadata repository allows you to more easily assess changes through impact and lineage analysis. The two ...