David Walker-278941 (4/22/2010)
sknox (4/22/2010)
Ah, yes, that does make sense. Still, it would be useful to know what the "units" are on the Time scale.
I think that with a lot of SQL Server work, the disk subsystem is the bottleneck... not just in these tests.
Sorry - just noticed the time question. The tests I did were just short 3 min. tests, and those "units" are just what MS Excel decided I wanted, and I didn't feel like fighting. Looks like they are about 15 sec. intervals.
Before you implemented anything, I'd suggest playing on a dev box with a copy of the database you are considering compressing with a more realistic load, and run tests for a lot longer.
We are just getting ready to run live with our first 2008 box (all 2005 otherwise) with a brand new app, so for now I'm just doing backup compression and not DB compression. From what I can tell it's just as easy to implement compression later as it is initially (keeping in mind a short burst of CPU/Disk use will probably be involved).
The Redneck DBA