A good example of calculations is currency conversion.
Suppose a cube is built out of data that is based on the local
currency.
Another cube has the exchange rates.
To view the numbers in a foreign currency the data in the local
currency must be multiplied by the exchange rate.
Hence a calculation is required.
How do the different OLAP platforms cope with this & what are the
strengths & weaknesses?
Presumably TM1 stores the calculations in RAM, so they are
precalculated ready for any queries - correct?
And Analysis Services does all its calculations on the fly, so they
are slower than TM1 - correct?
What about Essbase? Oracle?
What is the best way?

This is a current weakness with Analysis Services.
The current version doesn't precalculate any calculations.
So, yes AS can be slow if they are complex and involve lookups to
other cubes.

Essbase does do precalculations and historically this has caused some
data explosion problems; maybe not such an issue now though?

Not sure about Oracle Express.

As of today TM1 and Alea are probably better than AS when it comes to
complex calculations of a financial type. But both cannot handle the
same data volumes as AS.

So there is still no perfect OLAP platform...yet!

Regards,

John Keeley

www.johnkeeley.com



dontcutandpaste (AT) yahoo (DOT) com (sean dixon) wrote in message news:<f4cd0180.0406020622.7bd0be38 (AT) posting (DOT) google.com>...

Hi,

I think that the more interessing differences between the current OLAP engines
are in the possibilities of write-back options.

You most descide on your requirements,
what OLAP platform will be the best for you,
or what kind of disadvanteges you'll pay,
cause of the very different license models ...


Regards
Thomas Weiser

ANTARES Information-Systeme GmbH
http://www.antares-is.de
http://www.dynasight.net