indexed views are usefull for reporting access (SQL access)
to create a cube its not required and its an overhead.
Its why the Real project don't use indexed views.

The cube (in MOLAP mode) contains all the data both the detail and the
aggregations.
If your cube is slow maybe your aggregations are not correctly designed. Try
to recreate or add aggregation in your cubes.

There is no relation between the SQL performance and SSAS performance in
MOLAP mode.
The only impact of the source database is when you load the cube, not when
you use it.
Well... if loading the cube takes 4h and you do this every day, in this case
yes, optimizing the database can help you.

If you use ROLAP cube (Relational based storage) in this case the source
database has a big impact for the end user access. But I think you don't use
ROLAP cubes.

Can you describe where and when you suffer performance issue?
Have you try to change your views into tables and use SSIS to fill these
tables before loading your cube?
How your aggregations are designed?


"MIB" <MIB (AT) discussions (DOT) microsoft.com> wrote