This content has been marked as final. Show 1 reply
It looks like you are writing about the physical tables and indexes in your data model. In this forum, OLAP means something different, specifically the OLAP option of Oracle, which involves building multidimensional cubes in an analytic workspace. The cubes are generally built using the data warehouse or star schema as a source to load the cubes.
You could use OLAP to build cubes based on your tables, but I think that would be jumping ahead a little.
To address your immediate performance issues, you may want to consider partitioning the large tables you described, and using local indexes. Since the most recent data is needed most, consider partitioning on time (perhaps by month or quarter).
Without knowing more about your system, that's the most advice I can give at the moment. Please describe your tables and indexes in more detail, and also describe the reports (or post the SQL) if you would like additional help.