Roll out aggregate performance.

Question

Hi, experts,
    now we get performance issues about rull up step, it take too much time. even I have already try to speed up it. (implement notes 778709)
    some question about roll up. anyone can tell me the detail steps when rull up job running?
  1. why rull up almost take almost same time when I roll up 1 request and multiple request ( 5 requests)?  I think roll up 1 request should be much faster than multiple.
  2. sometime, the later roll up job will cancel the previous roll up job if it start on previous job still running. why?
  3. anyone can tell me how to reduce rull up time?
  4. is compress request in cube can reduce time for rull up ?
thanks a lot

Answer 1

Hi Jie,
The roll-up adds newly loaded transaction data to the existing aggregates. When aggregates are active, new data is not available for reporting until it is rolled up. For information on aggregates, see below. The time spent for the roll-up is determined by the number and the size of the aggregates; if aggregates can be built from other aggregates, they are arranged in an aggregate hierarchy.
Take the following hints into consideration in order to improve the aggregate hierarchy and, thus, the roll-up:
•     Build up very few basis aggregates out of the underlying InfoCube fact table
•     Try for summarization ratios of 10 or higher for every aggregate hierarchy level
•     Find good subsets of data (frequently used)
•     Build aggregates only on selected hierarchy levels (not all)
•     Build up aggregates that are neither too specific nor too general; they should serve many different query navigations
Monitor aggregates, remove those aggregates that are not used frequently (except basis aggregates)
I hope this information will resolve your doubts.