Intelligent cubes:
Intelligent cube is a caching mechanism which will allows the user to store the data in a micro-strategy environment.intelligent cubes are in-memory copy of data, which is placed in the intelligent server memory.
Features of Intelligent Cubes:
Intelligent cubes can be created for a number of performance reasons in microstrategy. Once built, you can create beautiful dashboards,reports and analysis off of an intelligent cube almost instantly. You can quickly analyse the data with out having to query your back and database. Here, the query will be hit the intelligent cube instead of DWH.
Drawbacks of Intelligent cubes:
Some of the features are not supported in the intelligent cubes. Consolidations are not supported in the intelligent cubes. So derived elements concept will be introduced. Prompts can not be included in the intelligent cubes.However, reports that access intelligent cube can be use prompts that access only data included in the intelligent cube. The cube data can be used in multiple reports. It always fetch the data from cubes.
OLAP services feature:
View filters, derived metrics and derived elements can not be included in the intelligent cubes, but these features are the main analysis tools for reports that access intelligent cubes.
Steps to create intelligent cube:
Go to public Objects -> Reports Folder -> Rt Click -> New -> Intelligent Cube -> Click 'OK'
How the Incremental refresh option works in Microstrategy 9.4:
Intelligent cube have a new feature to update the information without republishing it. It is called 'Incremental refresh'.There are different ways each incremental refresh option works:
How to define incremental refresh in report:
Microstrategy 9.2.1. introduced a new feature 'Incremental Refresh Options', which allow intelligent cubes to be updated based on one or more attributes by setting up incremental refresh settings to update the intelligent cube with only new data. This can reduce the time and system resources necessary to update the intelligent cube periodically.For example, If a user has an intelligent cube that contains weekly sales data, the user may want this intelligent cube to be updated at the end of every week with the sales data for that week. By setting up incremental refresh settings, he can make it so that only data for one week is added to the intelligent cube, without affecting the existing data and without having to reload all existing data.
Users can select two types of objects for the incremental fetch: a report or filter.
Filter: The data returned by a filter is compared to the data that is already in the cube. by default, the filter defined for the intelligent cube is used as the filter for the incremental refresh.
Report: The results of a report are used to populate the intelligent cube. By default, the report template used is the same as the intelligent cube's template.
When we run the reports
created in Microstrategy, they fetch the data form the warehouse to apply the
calculations and generate the report. When multiple users requests for the same
report but with different range of values or different filer conditions, then
the warehouse has to repeat similar calculations for each of the report and it
hits the performance.
To avoid this,
Microstrategy uses intelligent cubes which is object sitting in the middle
layer between reports and the warehouse.
The following diagram
depicts pictorially, the role of the intelligent cube.
Generally, intelligent cubes can be used in
performance cases. Example: Dashboard performance.
Most of the cases, In dashboards will take the
more execution time since its having multiple reports. Let assume, One
dashboard having 4 reports, If this dashboard tries to run, automatically the
existing 4 reports will be executed. So this dashboard will be display the data
after completion of these 4 reports.
If I create intelligent cube for this 4 reports,
then its fetch the data from cube which exists inside of intelligence server.
This is memory area. If a dashboard accessing the data from intelligent cube,
automatically dashboard will get the data from cube if the execution data
suites with available cube data. Here it does not need to hit the data
warehouse directly always.
