Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
goodsolution
Starting Member
38 Posts |
 Posted - 2009-05-16 : 09:49:54
|
| If i have 200GB of Cube,1.Generally how much time it will take to reprocess? 2.How many partitions i can do for that cube in SSAS?3.How should be the storage settings? (Ex: MOLAP or ROLAP or HOLAP) which one would be better?4.How often these partitions need to be processed? (Ex: daily?, weekly? or monthly?)-Thanks |
|
|
savior faire
Posting Yak Master
194 Posts |
Posted - 2009-05-26 : 10:10:45
|
| None of your questions can realistically be answered without knowing anything about your business requirements.One item to note that would be a general design approach, you can/should set up your fact tables in partitions based on the activity in them. For instance, your transaction fact tables could generally be partitioned by year. Prior years would not have activity, but the current year would.The amount of archived history stored in the cube could also be restricted to a couple of years. Analysis on old data may or may not be valid, since old data may be stale. Again, this would depend on your business and user needs.As far as processing a cube, this would also depend on the requirements for latency. Typically 24 hour latency and processing is suitable for many business situations. For example, in sales history analysis systems, up to the minute sales data is essentially of little use. Sales, marketing and finance would require a full days worth of sales rather than minute to minute activity.Hope I have provided some items to think about.Talk sense to a fool and he calls you foolish. |
 |
|
|
|
|
|
If My cube is 200GB of size?