No of Virtual and Concurrent users

[Guest Sambit Kundu]

Hi,

I am new to this tool.

May I know what is the maximum numbers of Virtual Users and Concurrent Users can be used for the WAFT tool.

And what is the Minimum Hardware specification we need to run the application in our system.

 

Regards

Sambit

[sergei 0]

WAPT does not have any predefined limit on the number of virtual users that it can generate. For different tests this number is different and depends on several parameters including the hardware of your system, network capability, tested server itself and, most importantly, the behavior of each virtual user.

 

The most important parameter is the delays between requests sent by each virtual user (this is also called "User think time" in the product). If you disable these delays, each user will send requests one by one very frequently. As a result you can easily achieve more than 100 hits per second even with 10 simultaneous virtual users.

 

That is why it is very hard to tell about "average" number of virtual users that can be emulated. If the performance of WAPT is insufficient for your test, you can use WAPT Pro. This version of the product can use several systems (load agents) for load generation. By adding more agents you can get almost unlimited load.

 

We usually say that we can generate about 1000 virtual users per agent. Actually this can be done with "average" delays between requests (10-20 seconds). If you disable all delays, you will only get about 200 users, even though they will make more hits than 1000 users with average delays. You will not be able to get more because of RAM limitation.

[Guest otana_uk]

Hello, also just getting started with this, so please excuse my ignorance..

 

For ~1000 virtual users, with request delays of 10-20 seconds, running from a single agent:

 

Would you be able to broadly advise on a suitable outgoing bandwidth requirement for this? Or are the actual sent requests not something that would trouble an average broadband connection (~10Mbps)?

 

I understand it depends on several factors - but is the outgoing bandwidth for a single agent something that needs to be particularly considered?

 

Thanks for any help!

G

[sergei 0]

I think required bandwidth = Number of users / delay * size of request * average request time.