Hi Ben, This is very handy info - thanks for sharing.
Kind Regards, Chris Mylonas On Wed, Mar 3, 2010 at 8:04 AM, Ben Gidley <b...@gidley.co.uk> wrote: > The benchmarks were done on our production replica which is > nearly identical to our live hardware. > > The key parts of the system is > - Oracle on a large Sun Ultrasparc box > - 4 active app servers - each quad core with 32Gb ram > - 2 active ZXTM load balancers/caches > > The load test is a mixed usage based on a complex model of expected user > journeys. The model allows for expected behaviour and caching on end users > browsers. > > Most our load is on the caches/db. The application servers so far are doing > very little (it would run off 2). We have them for later phases where we > will be adding more interactive content. > > > Ben Gidley > > www.gidley.co.uk > b...@gidley.co.uk > > > On Tue, Mar 2, 2010 at 8:56 PM, Kalle Korhonen > <kalle.o.korho...@gmail.com>wrote: > > > On Tue, Mar 2, 2010 at 12:13 AM, Ben Gidley <b...@gidley.co.uk> wrote: > > > The site has seen a lot of visitors over its launch period (I am not > > allowed > > > to release actual stats) but Alexa shows us just as the 1200 most > popular > > > site in the UK. Our load testing has shown the site scaling to 33 > > > pages/second (the transactions/second is a lot higher if you include > > > images/javascript/css). > > > > What's the benchmark system (mem, cpu) and is this mixed use with > > database operations? Certainly you cannot extrapolate the results, but > > just to give a rough idea of whether the test was done on a high-end > > cluster or throttled down lab environment. > > > > Kalle > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > > For additional commands, e-mail: users-h...@tapestry.apache.org > > > > >
