Forum Discussion

DeepaPai's avatar
DeepaPai
New Contributor
13 years ago

Maximum requests that a server can handle in parallel

Hi,

I am new to the web development. I am trying to implement web-services. I am using Amazon EC2 for storage.

The server that I am using should be able to handle 'n' number of requests per second in parallel. Because 'n' different clients can try to use my web-service at the same time from different devices.

How do I know the maximum number of requests that my Server can handle? I need to upgrade the server if it cannot handle the requirement. How Load UI will help me in finding this?

Please help me.

Thanks and Regards,
Deepa

3 Replies

  • SmartBear_Suppo's avatar
    SmartBear_Suppo
    SmartBear Alumni (Retired)
    Hi!

    I will from now on refer to Transactions Per Second as TPS. TPS is often a better way of measuring server performance than requests that a server can handle in parallel, since it implies that the requests are actually completed, and within a certain amount of time.

    First of all, we must remember that knowing the max TPS is not very useful without also knowing the response time (often called Time Taken in loadUI). For example, handling 200

    What you want to connect a Fixed Rate Generator to a soapUI Runner (or Web Page Runner depending on your needs), and plot TPS and Time Taken in the Statistics Workbench. You can then adjust the rate of the Generator to see how much your web service can handle.

    Since you're using web services, you probably want to start by creating a functional test for them in our other software, soapUI, and then load test that TestCase in loadUI.

    Here's a few pointers to get you started:

    Getting started with loadUI
    http://loadui.org/Getting-Started-with- ... -test.html
    http://loadui.org/loadUI-Demo-Movies.html

    Load testing soapUI TestCases in loadUI
    http://loadui.org/Load-Testing-soapUI-T ... -component

    Showing results / statistics
    http://loadui.org/Getting-results/getti ... d-new.html
    http://loadui.org/Getting-results/getti ... -view.html


    Hope this helps! With a little patience you should be on the right track soon enough

    Henrik
    SmartBear Software
  • DeepaPai's avatar
    DeepaPai
    New Contributor
    Thank you for the quick response.

    Would you please explain me this analysis with example?

    i.e. suppose I want my server to handle 500 users at a time, how do I test if it is capable using Load UI?
    I have given my web-service URL as web page address while creating the project.
    1. What should be the number of requests per second? [Is this 500/sec?]
    2. For how long should I run this test? [Minimum time duration for which we need to run the load test to evaluate server]
    3. Is ramp-up required to evaluate the server?
    4. What does discarded and failure count means?
    5. How do I know when the server has crosses the limit of maximum number of user that it can handle?

    Thanks and Regards,
    Deepa
  • subodh286's avatar
    subodh286
    Occasional Contributor
    Hi Deepa,
    Did u get any solution to your questions listed above? I also have similar questions around load testing of service operations using SoapUI/LoadUI.

    Please advise if you got any answers.

    Thanks,
    Subodh