May 12, 2008 at 2:46 am
Hi there! I have a database's table that have about 10M records. Users retrieve this data for viewing or getting statistics information based on this table. The question is: in such number of rows what more effective: to retrieve data for viewing and getting statistics from single server or to use replication and retrieve data for viewing on original server, but statistics from subscriber server? What more faster? I am asking because the data retrival for viewing has bigger priority and need to be done very fast, but statistics are also important. Thanks
May 12, 2008 at 3:56 pm
the fastest way to return data is to make sure the table is in cache, there is no contention, there are indexes which optimise data retreival and any updates. If the table functions as a heap, no updates, make sure the clustered index is not fragmenting.
Replication works fine but ultimately from memory is best, if from disk then lots of spindles is good.
[font="Comic Sans MS"]The GrumpyOldDBA[/font]
www.grumpyolddba.co.uk
http://sqlblogcasts.com/blogs/grumpyolddba/
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply