I like the article. The explanation is clear.
One subject I would like to get more info on is the remote processing of a partition. Right now, I process the cubes on a back office server and I push the processed cubes to an user facing server.
Using the remote processing from the user facing server may be a better option? I just do not want the users to be affected by the cube processing which is resource intensive.
Another article I would like to see is how to use SSIS to manage all these dimensions and partitions processing. Do you know of an article like that?
For the little story, I maintain a tally table for my partitions so I never have to manually add a new partition for a new time window. The tally table is a lookup based on calendar, it has a key,a lower and an upper bound. The bounds values gets changed as needed. The time horizon is not expanding, in this case, it is a rolling window. For old data we use another cube.