MOLAP is a type of on-line analytical processing methodology that is quite popular in the business intelligence world.
The “M” stands for multidimensional. From a technology standpoint, it basically means that the data is stored in a “cube” like fashion. The software takes the datasource and transforms it into the necessary data structure for the technology to handle queries, analysis, etc… The transformation typically does all of the necessary calculations and pre-aggregations so all of that data is available quickly for end users. Ususually little or no additional SQL would need to be run. In some instances, the data may be stored in-memory, which can also aid performance as that is much quicker than running from a disk, and no additional data warehouse usage is necessary.
In the right context, this technology can be an excellent addition to a BI implementation. Generally, they are best for implementations for smaller data requirements (under 100gb) and a reasonable amount of dimensions.
This has several advantages:
Calculations can be done ahead of time to speed performance when querying
Aggregates can be pre-calculated
The structure can be smaller than a corresponding RDMS structure in terms of disk space
Excellent for data sets with only a few dimensions
However with this technology come several disadvantages:
A structure to pre-process the data and calculations can be time and resource intensive
Datasets with many dimensions can rapidly drop performance
Data with high numbers of transactions can be slow to query
Several of the major vendors have a MOLAP technology available.
Cognos TM1 – One of the leaders with this technology. It’s proven especially popular in the planning and forecasting fields.
Microsoft Analysis Services – Along with Cognos is one of the more popular cube options.
Several other smaller vendors including PowerOLAP have had products in the past.
Palo is an interesting open source option that offers a multidimensional component
Copyright 2011-2018 BusinessIntelligenceBase.com