Embedding a cube-organized materialized view (cache, pre-computation tasks) for your multi-dimensional reporting and analysis might deliver a level of performance that would require an orders of magnitude larger hardware to satisfy the same queries using parallel execution and the detail data records.