View Full Version : Scalability and Performance

12-16-2008, 07:16 AM

I am an employee of one of the huge web portal company and we are evaluating BI products for our Reporting and Anaylysis prupose.

Anyone has any information on the following:

Who is Pentaho's biggest customer w.r.t., implentation size of its Reporting/Analysis/Dashboard products?

What is the maximum number of concurrent users the customer has and how they perform?

How good is the performance and scalability of the Reporting/Analysis/Dashboard products assuming that the server and other H/W infrastructure can scale to any extent?

P.S.: I had also sent a mail to Pentaho's sales team requesting for the above information, since I haven't got any response yet I am trying to find if anyone in the forum has any information on this.


12-16-2008, 07:39 AM
As far as "public" cases are concerned I think the most interesting case must be Mozilla. I recently saw a dashboard built on top of the Pentaho BI server with Mondrian quering Vertica and it looked absolutely great.
Mozilla's plan is to make that dashboard/Mondrian/query solution available to the general public, that's how fast it is.

Even though Mondrian can scale to large data sets with aggregate tables (see also: Pentaho Aggregate Designer) in this specific case, Vertica delivers a lot of horsepower to make it snappy.


12-17-2008, 06:30 AM
Hey, thanks Matt :)

I guess it's also relevant to say that the Mozilla project's front-end is built entirely using CDF - the latest versions are becoming tightly integrated with mondrian

12-24-2008, 04:23 PM
Can someone share some details about a sizable deployment in terms of number of rows for the largest fact table, and total size of database in ROLAP or MOLAP? We are looking for alternative to MS's analysis services and SQL server DW where the fact tables are in the range of a few hundreds of million rows to start with. Anyone can share some performance data?

Thanks in advance.


01-06-2009, 12:46 PM
Mondrian is ROLAP. With the use of aggregate tables it can be made to scale to those database sizes.
In that area we just released an aggregate table designer (http://wiki.pentaho.com/display/Surfboard/Pentaho+Aggregation+Designer+User+Guide) for Mondrian.