Hi

I am in the middle of trying to build a web based user interface to some
fairly large OLAP cubes (AS2K Sp3 - 32 bit). The cubes are somewhat
optimized and can generally produce query results in a decent amount of
time. The cubes are relatively big (25+ dims, 40 measures) - and its
impossible to reduce their size.

Up till now, I have been developing my OWC (10) app using the Foodmart 2000
demo DB, and its cubes - and performance has been great. I have now finally
started the migration to production, and found that everything grinds to a
halt.

For example, querying the sales cube over http, I get results and output. A
simple query against one of my larger cubes, produces no result:
"Failed to get data. If available, errors returned from the
provider are listed below."

I connect to the cube with the following statement:

Provider=msolap.2;Initial Catalog= databasexyz ; Data
Source=https://myserver.com ; Client Cache Size=0;Auto Synch
Period=100000;Execution Location=3;Large Level Threshold=100; Connect
Timeout=120;

As you can see, I have already tried to tweak the connection string to get
better results. If I use the datamember property, I connect successfully,
and I can recreate the same query I originally wanted to load using the
commandtext property. I know the commandtext property works, because I have
been able to test it using server side ADOMD.

So, I can only put it down to performance issues. Does anyone have a similar
experience or ideas to circumvent the shortcomings in either AS2K, OLEDB or
OWC?

thanks
Avi