Metadata and data warehouse

These are two key concepts in every BI project with MicroStrategy so we'd better have them clear since the beginning.

Metadata is data about data (what?). Ok, let's say that in the metadata you won't find any customer names or sales figures or any other fact, you'll just find information about the columns where this data come from, the datatype, and several other details about how to create the customer entity, for example, how does it relate to other elements in the project and where it is stored, plus the definition of all the reports.

While the data warehouse is the real source of the facts and dimensions that you will analyze. So, in principle, metadata and data warehouse are different databases and/or different instances; they may even reside in different hardware (and usually are, for example, in cluster production environments).

It is easy to deduce that the metadata is relatively small compared to the data warehouse, which is usually a huge, daily refreshed, historical storage with lots of records.

So AdventureWorksDW2008R2, which we just created, will be our "huge" 78 MB data warehouse; while we still do not have a storage for the metadata. Given that we are in a single-user training environment, we can use the same SQL Server instance for the metadata, but we'll create a new database to hold it.