In general, the more time your app spends waiting for external resources (database, LDAP, web services, etc.) the higher your concurrent request threshold should be (because requests will be consumed with waiting, leaving your hardware to do other things). But again, the only way to tell for sure is to actually test. Note that you don't necessarily need identical hardware to do useful load testing, it just has to be representative.
Also, you should check the network segment between your app servers and your database servers. If that's not across a dedicated private network, you might want to look at that as well. If you're sending a lot of data between the clusters, that might become a bottleneck for you if you're not running dedicated GigE. cheers, barneyb On Thu, Mar 5, 2009 at 8:25 AM, Kris Jones <[email protected]> wrote: > > Thanks Barney & Wil. Our application performs under extreme load 24/7. > Interacts with a database that is highly volatile (lots of inserts), > with tables/views containing millions of rows. Currently, the > instances are set to the default of 25 (2 of the stacks are set to 28 > per instance). I don't know how/when these values were determined, but > we're looking at it now, and want to understand how changing these > _could_ affect performance, either negatively or positively. > > Cheers, > Kris -- Barney Boisvert [email protected] http://www.barneyb.com/ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to date Get the Free Trial http://ad.doubleclick.net/clk;207172674;29440083;f Archive: http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:320143 Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
