KWymore (12/17/2013) Ford Fairlane (12/17/2013)
Koen Verbeeck (12/17/2013)
I have been enabling page compression on our entire data warehouse recently, so this was a pretty easy question :-)
Whats your performance been like since ?
I was about to ask this same thing. I have been considering doing this as well. So far I have only added page compression to our archive tables and not to ones being accessed more frequently.
If the compression brings you over 70% storage gain, than you should compress without any measurements, otherwise you should make some measurements concerning the CPU usage on compressed and uncompressed tables.
Once I did some investigation on compression on a small system. I compressed the entire database (mixed: page and row) and run a tool that was processing data against compressed and uncompressed databases. The processing time was same, and i obtained a storage gain. The avg CPU usage increased but not for much. It's recommended the avg CPU should not be more than 80%. So you can even compress the more frequently accessed tables.
If the reads operations dominate over the writes (e.g. more than 95:5) than you go with compression.
Igor Micev,SQL Server developer at Seavuswww.seavus.com