Submit and vote on feature ideas.

Welcome to the new Parasoft forums! We hope you will enjoy the site and try out some of the new features, like sharing an idea you may have for one of our products or following a category.

Load Test distribution models and you!

Thomas Moore
Thomas Moore Posts: 82 admin

Hey everyone! When running performance tests against services, it may be necessary to run a number of different distribution scenarios against the service in order to make sure that it is robust. Here, we will look at the different hit distribution strategies and when they will be used.

Steady load

This distribution type will run a constant number of tests against your service at any given time. This is to make sure that a service is able to maintain steady performance while being accessed by a static number of users over long periods of time.

Linear Increase

This distribution will begin running tests against your service with a low number of users, but will increase that number of users in a linear fashion as the test continues. This test is most often used to reach extremely high numbers of users at a given time to identify break points in either the network, service host, or the code within the service being tested.


Similar to the linear increase, this test will simulate a small number of users at first, but the exponentially ramp up to a high number of users at once. After the test reaches this peak, it will ramp down in a similar fashion. This test type is often used to test break points in the scalability of software, making sure that it can spin up threads quickly to accommodate the large number of users coming in at the same time, then dropping those threads properly once the tests have finished running.

Buffer Test

This is the most unique of the scenarios for testing that is offered with Load Test. This test will simulate no tests being run against the service at first, and then immediately jump to a large number of users with no spin up time. After a period of time, those user hits will immediately stop, only to wait more time and start up again. This test is used to simulate some services where it is expected that it will receive a large number of hits for short periods of time before going back on cooldown. An example of this would be concert or convention tickets going on sale, and having many users log in to try and get their tickets before anyone else.