Three-layered Meta-data Management:
Gather-Store-Deliver!! Every cent spent in IT is to accomplish one or more the three data related activities. meta-data implementation should have a clear concentration on one of the three layers at the time. For example, if the meta-data is not captured, there is no point to allocate lots of resources working at the delivery layer. For each layer, standard tools and access methods must be defined. The integration could be done at any or all layers depending on the priority of the organization. Click here or the picture to see the example diagram of “gather-store-deliver” for data warehousing.

Gathering meta-data
For example, in creation of meta-data for a data warehouse, there are different categories of work. At the gather layer, the organization should capture all the information and details about data extraction, filtering, validation and cleansing, merging or splitting, aggregation, transformation, loading and updating, and archiving. At this point, how to store the meta-data is not the focus. They may remain in the ETL (extraction, transformation and loading) tools, the CASE tools such as AllFusion ERwin Data Modeler^™, data quality tools such as Vality, or just stored in an textual document a spreadsheet.

Storing meta-data
At the store lay, the meta-data work focus on how to internally represent the data structures, schemas and storage mechanisms that are required to persist the meta-data. Many organizations choose to purchase a commercial meta-data repository product to integrate the captured meta-data. Others build their own. The decision is critical to the success of meta-data initiatives. Again, a dollars spent on capturing meta-data is not going to be available for storing meta-data. One has to evaluate the business case and technical situation to craft out a solution. Yet, at the store layer we are not concerned with how to get the information to the hands of data customers.

Delivering meta-data
Delivering meta-data to the users is the third critical layer. It is the type of work that will attract users and possibly funding for the meta-data project. Technologies used include Internet and Intranet, Web portal, knowledge Management, document management, groupware, project management, etc. In many cases, meta-data that are not integrated at the store layer may be visually integrated at the delivery layer by designing a well-linked, easy-to-navigate web front-end. It is important to have different delivery methods and front-end designs to fulfill various needs of different audiences.