Dynamic horizontal partitioning

I have a database which requires merge replication which can be horizontally
partitioned into 25 branches.
Would this create 25 distribution databases and is there a high impact on
the performance on updating these databases?
Or is the best solution to use is dynamic horizontal partitioning?
Thanks
mm,
there is no need for more than one distribution database. Your options are
to create 25 separate publications and a distinct subscriber for each or (as
you say) you can use dynamic horizontal partitioning. I would favour the
latter as it is less administration, especially if any schema changes are
needed. To implement, you can easily add -HOSTNAME as a merge agent
parameter for each subscriber, and use this to filter.
Regards,
Paul Ibison

DW partitioning advice needed by lost soul

We are putting our first DW together. One of the consultants, not familiar with the MS SQL database, suggested we consider "parititioning" our data for ease or reloading, distaster recovery etc. His idea was that we could parition the data based on, say, year, and that way, in case of a disaster, we would only need to reload that year's worth of data from our backups. (So I guess he is talking about software partitioning, not hardware partitioning?)

I can't imagine how this approach wll work with our nightly backup strategy or the changes the need to be made to the DW import code to to handle these partitons. Furthermore, in case of a disaster how is one to even know what parititon(s) to reload. (I'd be incliined to reload the entire warehouse to make sure I got everything!). I'm lost. Any thoughts you can share on this headache will be much appreciated!

barkingdog

Not sure if you are asking question on the right place. This forum is specifict to issues related to Analysis Services only.

Edward.
--
This posting is provided "AS IS" with no warranties, and confers no rights.

|||

My mistake -- I will contiue to look for the the DataWarehousing newsgroup!

Barkingdog