In a word, yes. There's lots of approaches you can take (commercial, open source, roll your own) but the concept is well proven. I routinely run tests at higher volumes than you mention with no problems. My personal approach is to run lots of jmeter processes as isolated silos and group the results later - this is logically very scalable. But there's been significant work lately on making the Distributed Mode more efficient so you may have success there.
In industry terms I personally know of several high profile companies that use JMeter as their standard tool and there is also a video kicking around where Google talk about how they use it for their load testing. In each case they opted to build their own framework; it's not really that hard and the cost savings are stupendous. On 3 Mar 2013, at 19:54, Sarndeep Nijjar <[email protected]> wrote: > Hi, > > Im looking to delve further into jmeter and use it as the standard tool I use > for all of my clients that require performance testing. Ive started looking > at running jmeter via cloud injectors so I can get some pretty decent load, > but I want to understand how scalable it is. > > Has anyone had any experience at running Jmeter, either via cload injectors > or just from a big server for say 6000+ concurrent users? did it perform > well, were the results accurate? > > I hope im not being too vague! > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
