Data Integration Metadata Management

You may like to refer the section Metadata Management** for some more subject matter expertise. Metadata is essentially "data, information and knowledge on data". A Data Integration tool not only has to have good capabilities to use metadata, but also to create it.

• Capturing of Metadata, while designing of the Data Integration routines: As you design your data integration activities and data structures, the metadata should be automatically create and revised as you go through the enhancements.

• Capturing of Metadata, while executing the data integration routines: There is design metadata and operational metadata. The DI tool should generate the operational metadata**

• Metadata mapping capabilities to propagate data definitions: A DI tool should be able to pick-up the metadata of the source systems and can use it as a base for creating metadata for the transformed/integrated data.

• Wizard to create Metadata: The automatic creation of metadata is typically limited to logical, physical and implementation layer**. A wizard should enable you to add conceptual and contextual layers. Secondly, it should also allow you to add metadata for logical/physical/implementation layers, which have not been automatically generated.

• Wizard to import and export metadata: This is a standard feature of most of the metadata layers in any kind of tool categories. One should be able to import and export metadata on full or conditional basis and through various scheduling modes.

• Supporting the central as well as distributed metadata repository: As mentioned in Metadata section, a physically single enterprise level metadata repository is generally not achieved and mostly organizations have distributed metadata repositories. A DI tool metadata layer should be able to work with both these metadata architecture scenarios**.