At Oracle OpenWorld, I attended a session by William Endress and Martin Gubar entitled "Using Cubes and Cube Organized Materialized Views to Enhance the Data Warehouse". Both presenters are Product Managers at Oracle. They talked about the Oracle Database OLAP Option. Basically, you can build OLAP cubes within your Oracle database. These OLAP cubes are embedded within the database. Cube data is stored in Oracle data files and cube meta data is stored in the Oracle data dictionary. You get all the security of your existing Oracle database without having to set up a new security infrastructure. Data does not have to be moved to another product. Everything is self contained.
Within Oracle 11g, they now have "Cube Organized Materialized Views". What this means is that your cube is like a traditional Oracle Materialized View and the Oracle SQL processor re-writes queries that can take advantage of the summary data in the cube. Access to summary data in the cube is fully transparent. After you create your cubes, your existing BI or analytical applications are automatically able to take advantage of the cube through the power of the Oracle SQL processor. You use the same apps you always used. You use the apps in the same way you always have. You just get better performance.
They also showed a case study comparing cube oriented materialized views vs table oriented materialized views. While loading and aggregation with cubes was faster by about 2-3x, query performance was faster by about 7x. This was significant. Very powerful product indeed.