There is no technical limit to the size of an OLAP cube. However, there may be environmental limits (disk space, memory, etc.) as well as practical considerations (does a cube really need 25 years of history, or is 5 years enough)? I would also ensure that large cubes are tuned properly with all of the proper aggregations to maximize performance.
There are very few limitations for SAS OLAP Cubes.
The only ones I come to think of is number of herarchies (128) and measures (1024).
In my opinion they are no real life limitations. I can't see an end user chsoing between thousands of measeures and hundreds of hierarchies...
Maybe you could track this to SAS support.
Some monitoring is available in the OLAP Server Monitor plug-in for SAS Management Console.
Also check the IP admin docs:
We recently got out-of-memory errors when opening cubes after upgrading from SAS V9.2 M1 to M3 and from Web Report Studio V4.2 to V4.3.
Turns out that we were unknowingly selecting ALL fields in the cube for display in Web Report Studio via the Select Data panel, instead of selecting say 2 or 3 dimensions and a few measures.
And so selecting all fields caused the OLAP server to generate a CROSSJOIN for all dimensions (say 25 of them) and all measures. With 250,000 rows, this blew the Java VM memory allocation every time, regardless of how much memory was allocated.
128 dimensions is a limit for financial services :smileycry:
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.