Is there an (open source) MDX-based OLAP server?

Question

I was wondering if there is in memory that MDX can handle.

I found a wikipedia article . But he says nothing about the functions in memory ...

The only thing I know about them is Mondrian, which is also one of the two open source solutions mentioned in the wiki article above.

So far I have worked a little with Mondrian ... I don’t think that Mondrian is a real memory OLAP server.
That's why:
It has a cache in memory that contains queryresults. But their first execution takes a lot of time and must be downloaded from the RDBMS. ( http://mondrian.pentaho.com/documentation/faq.php#Scalability )
I think the best way:
Load all fact and dimension tables into Cache, and then run each query against that data in memory.

AFAIK, Oracle will release with the release of corporate release 12c this year the ability to have some (or all) tables in memory. This will speed up OLAP servers that simply query RDBMS fact tables with SQL.
... but the corporate edition is very expensive ...

I would like to hear some other opinions about this.

Yours faithfully,
Dennis.

Answer 1

Look at the Sparkube project that provides any Apache Spark dataset as an OLAP cube that you can query using MDX from an Excel PivotTable table, Tableau or ActiveUI software, the web interface that Sparkube uses. Aggregations are distributed in the Spark cluster and performed on the fly in memory.

http://activeviam.com/sparkube

The Sparkube MDX engine is one that was developed by ActiveViam for their commercial analytic platform, so it has been used and tested by many.