ElastiCubes

ElastiCubes are high performance data stores optimized for fast query response.  ElastiCubes allow you to bring in data from multiple sources, merge, manipulate and then query it as if it was one consolidated data set.  ElastiCubes perform so well that in most cases creation of dedicated OLAP cubes and/or optimized data marts are completely unnecessary - even when dealing with millions of rows of raw data.

Conceptually, an ElastiCube is just like a big table.  It is made up of fields where each value in one field has a corresponding value in another field.   The data for an ElastiCube can come from one source, multiple sources or even from multiple phyiscal locations.  Once the data is inside the ElastiCube, it is all the same and every field coming from every table can be analyzed in the context of any other - quickly.

For example,  the image below defines an ElastiCube with 24 fields.  The data for the ElastiCube originates from three tables: PurchaseOrderHeader (11 fields), Employee (11 fields) and EmployeeAddress (4 fields).

 

You may have noticed that the total number of fields at the source is 26 yet the number of fields in the ElastiCube was said to be 24. This is because the three tables have a shared field.  The three connected lines, called relationships, indicate that EmployeeID field - which exists in all three tables - is in fact the same field.  Relationships allow ElastiCubes to correlate between data that originated from different sources.

 

ElastiCube Advantages

ElastiCube technology make queries over millions of rows of raw data return in seconds, with moderate hardware requirements including standard desktop-class computers with a reasonable CPU and 2 GB of RAM.  More importantly, ElastiCubes can do this without having to pre-aggregate and pre-calculate the data ahead of time and store it on the hard-drive, thus radically reducing required import/processing time and storage space.

 

ElastiCube Use Cases

ElastiCubes are most useful when one or more of the following is true:

1. Large amounts of data need to be analyzed
2. Data for analysis originates from multiple disparate sources

 

Building and Deploying ElastiCubes

Once an ElastiCube schema has been defined, the ElastiCube can be built on a local machine for personal use or on a server station for use by multiple users and/or applications.  The build process imports data from the specified source tables into the ElastiCube server where some addition internal processing is done before the ElastiCube is ready to service query requests.

See Also