We have a cube that is 40 gigs.
We have a lot of data and very few dimensions.
We have performance issue on Read Disk queue lengh.
How should we setup the properties of the Analysis server to have the best
performance ? ( process, threads, etc.)

Thanks.

In general, you're going to need to experiment to see how to improve the
performance, a good guide for this is:
http://www.microsoft.com/technet/pro.../ansvcspg.mspx

There are quite a few properties to be talking about as well as design
considerations. While you have few dimensions, are you measures distinct
count or aggregate (count, sum, etc.) in nature? What type of queries are
you trying to improve - ones that involve caclualted members or simple
pivots?

--
HTH!
Denny Lee
<dennyglee_at_hotmail_dot_com>

Blog at:: http://spaces.msn.com/members/denster/



"Ben" <Ben (AT) discussions (DOT) microsoft.com> wrote

We use count, sum, avg, min ,max, etc.
We use queries that involved caclualted members and simple pivots

"Denny Lee" wrote:

have you created partitions?
have you used the usage based optimisation wizard?

do you use AS2000 or 2005?

"Ben" <Ben (AT) discussions (DOT) microsoft.com> wrote