At the moment, at my company, we're facing problems with an increased duration in cube processing time; from 3.5 hours to 6.5 hours in one week. This could be due to increased activity with SQL databases on the same box. I want to find ways of reducing the time the cube takes to process (last resort - moving the cube to another server)
We have standard edition and the cube has 3 partitions (the limit of standard). So, incremental processing is out the question. I guess I have a couple of questions. Is there really any benefit of having partitions on Standard edition since there's no parallel processing; I'm assuming it speeds up query processing time!?
The other question I have is around types of processing. Our cube has dimensions that have new data on a daily basis and of course so does the fact tables. The dimensions have updates to existing rows. The cube is Process Full each night.
Is it still valid for me to perform ProcessData on the cube each night instead of doing a ProcessFull? The latter seems excessive but I want to make sure I capture new data and ALL updates to existing dimension rows if using ProcessData.
Any help is much appreciated.