Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am trying to figure out a practical way to stress test a server for 300 to 600 simultaneous connections. Any advice?

Thank you everyone for the help. To be more specific (sorry I wasn't before) this is a Flash Media Server on AWS that will be streaming live video. I've been having problems with the video freezing/buffering for everyone and I need to verify if its on the user end, upload end, or server end. I mainly need help with stress testing the server with 300-600 multiple request before going live.

share|improve this question
What kind of server, and what kind of application on the server? I could guess (and I will in an answer below), but we'd rather know. – Joel Coel Jun 20 '11 at 21:39

In the generic case (any application, not just a website) you can do this with expensive software like HP LoadRunner. If it's a website, though, you've got a lot of low-cost or even free options as well, assuming you've got enough client horsepower to drive the load. The benefit of more expensive software is making it easy and quick to change things like percentage of sessions that do CPU-heavy searches vs just browsing (for example.)

share|improve this answer

Assuming you're asking about a web server running IIS/ASP.Net, Microsoft supports this through a load test feature in Visual Studio Team System.

share|improve this answer
Or VS Ultimate Edition! – mfinni Jun 20 '11 at 22:46

You might want to look at a 'hosted stress test provider' like BrowserMob. I don't know much about them, other that their name comes up regularly, and their website seems to list "Flash Video Streaming" as a supported feature of their load test service.

IIRC, BrowserMob's test agents run on Amazon EC2 too. So in order to get a realistic test, you might want to be sure that a large portion of the agents come from a different EC2 Availability Zone -- so that you're testing with actual WAN traffic.

share|improve this answer

300 52kbit streams would be something like ~1.5megabits of bandwidth. Have you considered a multicast streaming solution which would reduce your bandwidth needs? Do this inside of EC2 and you are streaming either into our out of the cloud and you will wind up paying quite a bit.

share|improve this answer
300kbit*52 = 15600 = 15,6mbit/s – pauska Jul 29 '11 at 11:18
Yep, that's what I get for math in head vs writing it out, order of magnitude off. Thanks for catching it. But what you have done is made my initial point so much better....who has 15 megabits of h coming out of their streaming server. If you get to this level then you are going to need to stream using a CDN proxy solution or you need a different streaming model. – James Pulley Jul 31 '11 at 13:22

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.