Question

I have a cube that I have build that has data across multiple servers. After the cube is deployed to the SSAS server, does it interact with the SQL servers that contain the initial data in which the cube was based on? The reason I ask is because I have potentially a lot of users and some of the data is on one of our production servers which we don’t want to be accessed during a query to the cube.

Thanks,
Ethan

Answer 1

A typical SSAS Cube copies all the data available to it (as per the tables/views you pull into the DSV) to it’s own location, you can validate this by going to the storage path as defined in SSAS Server options and looking at the folder sizes. When you query the cube, it will use this ‘copied data’.

Having said that, there are exceptions:

If you have ROLAP dimensions it can go through to the underlying data:

http://technet.microsoft.com/en-us/library/ms174915.aspx

If your cube is set up for proactive caching, then it could query the underlying databases itself in order to stay up-to-date:

http://msdn.microsoft.com/en-us/library/ms174769.aspx

Those are the only two I’m familiar with.

Do bear in mind that deployment will generally require processing afterwards, unless you’re restoring from a backup you’ve processed elsewhere. Also bear in mind at some point you’ll probably want to add new data into the cube, which you say comes from the production databases you don’t want to interrupt.