An OLAP cube is one of the technical components of on-line analytical processing. It’s a way for data to be stored for access by in-memory or other BI analytical functionality.
The name originates from the idea that the data is in a multidimensional state. From a visual standpoint, its often represented as a cube, compared to the table look of a traditional database structure. Often times there are more than three “dimensions” used, but the concept is the same. People from the Microsoft Excel world often conceptualize it as a multi-layered spreadsheet.
If you think of the data in this regard, the speed of this type of structure for queries is apparent. Rather than hitting several different tables in a typical SQL query, in a cube all of the data and hierarchies are stored together. In addition, often the calculations and analytics can be pre-calculated ahead of time, saving processing time when you are running a report.Another performance consideration is that oftentimes these can be placed “in memory”. This means that the data is more quickly accessible from a data read standpoint, as compared to a standard disk based database that can be much slower.However, creating these structures are not as simple as flicking a switch on the database. In fact most databases don’t have this functionality built in. It’s a BI focused ability and is usually driven by the software.
Cubes also need to be created and built. Loading the data and dimensional calculations takes time, and it needs to be considered ahead of time for the various reports and dashboards that will be built off of it.
One drawback for this capability is that your BI package needs to have this OLAP functionality available. Many of the big BI vendors have cube technology within their BI offerings. Hyperion has Essbase, Microsoft uses SSAS, and Cognos also has a cube functionality.
Copyright 2011-2017 BusinessIntelligenceBase.com