Delta cache or compression

Question

Hi,
I am now investigating delta cache.
if data are compressed on the cube, delta cache cannot be used. And I have heard compression is still recommended and delta cache is less important than compression.
But I think it is best when only older requests are compressed and leave some request not to be compressed on the cube,if delta cache still work, it is best to do so (and then aggregations should also have request for delta cache).
if something wrong ore there are other good way, please point them out. (excuse me, for some reason I cannot be test now..)
Kind regards,
Masaaki

Answer 1

<b>Delta cache-</b>
The delta process enables individual data from a query to be inserted into the cache. In this way, when they are changes to the data basis, the complete data for the query no longer has to be read from the database again. The results of a query are thus independent of whether the data is read directly from the database or from the cache.
A complete restructuring of the cache is always necessary, however, if metadata, master data , hierarchies, exchange rates, etc. change.  But a complete rebuild of the cache object also has to be performed for InfoProviders that are not suitable for the delta process when there are changes. InfoProviders are delta process enabled when a time stamp, sequential number, or, as with InfoCubes, the request shows which entries have been added.
<b>Dependencies</b>
If the query is defined on a MultiProvider, you can still optimize the query runtime when you set the partitioning type accordingly.
Compression is still recommended and you cannot escape doing it ,based on the accuracy and reconciliation of your data for the period you should set up your compression strategy as once the data is compressed you cant delete (requests)it.
PS- However,you can delete compressed data using option SELECTIVE DELETE.
The typical measure taken in such situations, is to build aggregates on these cubes (instead of compression). This will not help, but only postpone the disaster to a later point in time. And at some point compression is not possible in a reasonable amount of time anymore, because deleting a partition takes so long that the whole workflow on that cube (and others later up the chain) is not fitting into the allocated timeframes.
Hope it Helps
Chetan
@CP..