Hi All, I am basically a working

[TDKID]

Hi All,
I am basically a working on a database and a reporting tool for our project we are analysing on useing SQL Server and MS OLAP where we need to know about the performance and architecture of these two with respect to Teradata and MicroStrategy.

It would be better it I can get some web links along with the reply

Thankx in advance

 

[MDXer]

Well the Analysis services (OLAP) portion if designed right will out preform Microstrategy. Microstrategy is a ROLAP (Relational Olap Technology) Meaning the OLAP Structure lives within a RDBMS, whereas Microsoft Analysis Services allows you to use MOLAP (Multidimensional OLAP Technology).

The Molap cube is seperate from the underlying fact records after the cube has been processed. The aggregations are created when the cube is processed. This is one factor in it's speed. Yes in ROLAP products you can have aggregate tables and lookup tables that aide in query preformance but these give you a preformance hit at processing time, and also increase the storage space needed.

Analysis Services not only aggregates the underlying fact records but allso allows you to preaggregate, or query optimization, the values contained within the Molap cube. and can identify and optimize based upon query history.

Space Requirements in Analysis services are smaller. Analysis services does not build NULL values into the molap structure meaning the the amount of data is less, where as in ROLAP and some other MOLAP products NULL Values are alloted space. I personally have seen 860GB built into a 100GB cube that was 250GB after Query optimization had been applied.

The fact that the MOLAP data is disconnected from the underlying fact data (Unless Drillthrough is allowed) also makes it easier for Server admins to do maintance on the RDBMS servers.

The baseline of what is best is dependent upon a companies needs.

HTTP://www.olapreport.com

has some good information.

As for the teradata i have heard a lot of positives about it but it is costly, and probably something a company would avoid unless the data volume warrants something of this level. You may find that MS SQL or ORACLE may provide the preformance needed.



"Shoot Me! Shoot Me NOW!!!"
- Daffy Duck