# Thread: Calculating the Real users...

1. ## Calculating the Real users...

Hi,
I am currently doing a Load testing and the scenario of my test is as follows:

One script with multiple transactions running for 30 min with 50 Virtual Users.

My question is, how to calculate the real time users in the above scenario.

Can some one help me out ...

Thanks,
Venkat

2. ## Re: Calculating the Real users...

Are you trying to figure out how fast to make the Virtual Users work? Or did you mean how to project a number of actual users based on the information you have about virtual users?

It's the same equation, either way. VUser (A) works at speed X, real user (B) works at speed Y, solve XA = YB.

There a few ways you could do this, and it could be argued you should do them anyways:

- Experiment with different frequencies and see what happens. Use these numbers to establish ranges.
- Go watch a user work and figure out how fast they work.
- Get a log file and parse it. Derive frequency from that.
- Make an educated guess.

Is that what you needed?

3. ## Re: Calculating the Real users...

[ QUOTE ]
solve XA = YB.

[/ QUOTE ]

This is not entirely true because both tests will produce different performance characteristics on environment.

4. ## Re: Calculating the Real users...

Venkat

Do you need to know the response time? If you need this, you must take the response time for each virtual user in each execution and then calculate the average with a percentil or a standar deviation of this measures. It is necesary to know the amount of executions that each virtual user did.
You must to know that the response time in test environment can be different to production environment.

Luis Cardona

5. ## Re: Calculating the Real users...

Hi Eric,
I am looking for how to project number of actual users based on the information that I have about virtual users.

Let's say I am ruuning the script for 50 users for 30 min. Total no.of transactions for each step in the script are 300(for example)...

Is it right to say the actual users are 300 because each step has been executed for 300 times?

6. ## Re: Calculating the Real users...

[ QUOTE ]
...Is it right to say the actual users are 300 because each step has been executed for 300 times?

[/ QUOTE ]
No and never unless you ran 300 users. It is correct to say that you ran 50 users for 30 minutes.

7. ## Re: Calculating the Real users...

Jake and ASH are both quite right. 50 users operating six times as fast does not necessarily equal 300 users. I am wondering why I was so flippant about it.

What if there was a software or hardware condition such as cumulative memory footprint, or number of database connections, licensing, or something else where a threshold would be reached by having 300 users on the system at once? You might certify a system to support 300 users that can't possibly support 100. You can KNOW unless you've actually had 300 sessions active on that system.

Now, presuming that your model has each user connect once in 30 minutes, you might be able to say that you have consumed a similar amount of network bandwidth, generated a similar number of database queries, or consumed a similar number of CPU cycles as 300 users, as a result of doing the work of 300 users. That can be risky; caching effects can be subtle, and load tests are already crude simulations at best in most cases, for a number of reasons.

When I write up these kind of results, I might make those connections, but I try not to assert that six times 50 is just as good as 300, for example. Sometimes you have to make the best use you can out of the information you have, but I recommend avoiding any kind of guarantee based on this information. This is more an educated guess than a calculation.

8. ## Re: Calculating the Real users...

[ QUOTE ]
One script with multiple transactions running for 30 min with 50 Virtual Users.

My question is, how to calculate the real time users in the above scenario.

[/ QUOTE ]

No you cannot calculate this with data you have given, talk to business for this.I can simulate the transactions thousands times with 1 users but that is not thousand users.Users and Transaction count is totally different theory.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•

vBulletin Optimisation provided by vB Optimise v2.6.0 Beta 4 (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.