I am testing a rolap cube with 1 facts record and some small-medium molap dimensions.
Drilling down these dimensions in MS is realy slow...
Changing the cube's storage to molap makes dimensions drill supper fast.
What mite be the problem when cube is rolap??
The MDX you send from your client is translated into (T)SQL and then sent to the database engine. That is much slower than using MOLAP. A good scenario for ROLAP is very large dimensions with millions of members that you cannot load into a MOLAP storage. Even in this scenario you need to add summary tables or materialized views to speed up ROLAP.