Thanks for the help

Getting my head around the architecture of data warehousing is proving to me tougher than I expected.

So basically, the data source is irrelevant when it comes to deciding between ROLAP and MOLAP as these options are only available once the cube has been created, and the only difference between them is the way the data in the cube is stored.

MOLAP - Multi-dimensional, creates its own db files based on the data source. - Correct?

ROLAP - Can you please elaborate on the topic of indexed views in this regard. Are the indexed views created on the data source or on the cube itself? In ROLAP, isn't the cube just a representation of the data source and not physically stored on the analysis server?

Thanks