Business Intelligence OLAP is a component of the Business Intelligence software suite. OLAP – Online Analytical Processing, is based on pre-populated,
highly aggregated multi dimensional cubes. These cubes can be used as a data source by a number of OLAP reporting and analytical applications.
The key feature of an OLAP cube is it high level of aggregation, providing fast OLAP analysis of underlying data and fast answers to user queries. Despite being aggregated, cubes can in fact be quite large in size due to the level of aggregation and more importantly, the number of dimensions held within the cube.
The reason for this fast response and potentially large size is that all possible combination of answers, based on the measures and dimensions being used, are calculated. This may be considered wasteful, particularly if only 20% of a cube is actually used. However, the fast response times to queries far outweigh any extra time it may take to populate the cubes with all combinations.
If the extra time required to process all combinations becomes a problem, a subset of the cube can be refreshed. The remaining ‘data’ if required, if then refreshed as required. This will have an impact on report response time. The problem here is determining a good subset of OLAP data to refresh.
The cubes can be developed in a number of OLAP tools. The most common is Microsoft SQL Server Analysis Server (SSAS). However, applications from Hyperion and Cognos can also produce OLAP cubes The latter calls its cubes Powercubes.
Depending on the cube size, they can be updated frequently, or for large cubes, refreshed overnight. However, frequent updates are only useful if the underlying data warehouse is also updated frequently.
There are many reporting and analytical applications which can use OLAP cubes as a data source. These include Microsoft SQL Server Reporting Services, Panorama Novaview, ProClarity, Performance Point Server and Business Objects, to name a few.
Cubes are widely accepted and considered excellent reporting sources due to their response times and level of detail.