Using ROLAP (Relational Online Analytical Processing) or MOLAP (Multidimensional OnLine Analytical Processing) is something that must be raised in physical design stage and affects the way the tool of exploitation of Data Warehouse 'attack' to the data .

If ROLAP, the tool will exploit the information directly attacking the stars that were created in the relational DBMS chosen.

If you use MOLAP technology, although we have created these stars, the tool will create their own multidimensional cubes optimized for the engine to implement. In this case there are enough deployment options: you can make it go directly from an ODS (Operational Data Store) to build MOLAP cubes, and also keep the stars in the relational, or even create cubes directly, bypassing the relational.
I am also in favor of retaining all implemented on relational environment, and not bind us and the technology implemented by each tool for creating cubes.

By way of example with SQLServer Analysis Services cubes implementing ROLAP, MOLAP or HOLAP is as simple as dialing a radio button to define the cube, the tool makes it transparent, so it can be as simple as considering what option is better depending on size and use and growth forecasts of DWH, do some test and choose the one that best suits our needs.

Anyway, the most common is that the tools opt for the use of one technology or another.For example, the DSS (Decision Support System) Microstrategy favors the use of ROLAP with a relational DBMS and SAP BW or Essbase Analytics (Hyperion), always implement their own multidimensional cubes.