top of page
Writer's picturetakery reddy

Software testing company in Texas

When should I begin load testing?

What is the best strategy for load testing invest in a realistic test or take a quick and dirty approach?

Many testers to realistic goal, but the establishment of a realistic simulation can be time and effort. This can significantly delay the tests, leading to serious risks. As Kent Beck and Cynthia Andres seen in Extreme Programming Explained, catching problems early cost less to fix the end of the development lifecycle.

I would say that in many cases, this approach gives better results. We can spend 20 percent of our effort typical test setup and still learn 80 percent of what we know and we can find problems when they are still cheap and easy to fix.

The 80/20 rule for performance testing

The 80/20 rule, also known as the Pareto principle states that for many things, 80 percent of the effects are derived from 20 percent of cases.

Creating a load test script can be a great example of the Pareto Principle in. Configuring a load test is difficult. There are tons of settings you can change to get the kind of user behavior and simulation you want.

Many people approach the load testing with the aim of perfectly simulate reality in their load test, so they will learn exactly how real users their production system will be able to handle.

Unfortunately, simulating reality is difficult, which means that it takes time and effort to develop such a test, and many times even a major effort fall short and miss a critical aspect that causes unrealistic simulation .

The end result can be an extremely costly and / or time load test which might not even give you the results you want. Organizations that spend tons of time and money on a load test are not likely to do any load testing, except when it is necessary.

Often, this will be just before the launch, when there is little time to correct any problems and the cost of fixing them will be high because the problems are at a late stage.

Simplify testing and start early

Here’s where the Pareto principle comes into play: if you do not try to simulate reality perfectly, your configuration will be much simpler. You can just spend only 20 percent of the up the load test development effort.

The result you get may not be a number that you say exactly how many users your site could manage the production, but it is likely to tell you if your performance is below target, and it will highlight your chances bottlenecks throttle.

Because the test is much simpler to set up and run, you can run several tests at different stages of development and learn much in total you have learned from a single test.

Being able to run other tests earlier to detect and fix performance problems earlier, and fixing a matter as soon as possible usually means a time-saving development, resulting in lower total cost of the project. More frequent testing will also result in a team that is generally more confident and involved in product performance.

Best practices for load testing

It was a lot of abstract ideas, soft, but how do you really?

The simulated user in your test could be much more static in behavior that a real user may simply access a page set or sequence content without randomization. Or you can have your simulated users access a randomly selected page on your site, choosing to ignore the fact that some pages are visited much more strongly than others.

If you are testing an API driven application (such as a mobile app), you may find it best not to worry about user flow and instead of trying to end-point is the exercise as a PLC mixture or perhaps even individually.

. Then do the same for each endpoint (you can exercise over multiple endpoints in the same test if you want).

The absolute easiest test you can do is hammer one end point of the API at a time. This is easy to setup, but if you run this test several times to each major evaluation criteria and at an early stage in the development of your application, you will be sure to come your release date of the performance.

Build realistic simulation

There is certainly value in the execution of more complex tests, but I would say if you are not being simpler execution, load testing and stress more frequently, then you should probably start there before going for large scale, complex tests that closely simulate reality. Simple tests are the result of low hanging here:

• Execution of simple tests means that each test requires less time and effort

• Less time and effort by test means you can run more tests

• Other tests give you more opportunities to iterate and optimize your test setup

• More and more small tests also allow you to run earlier in the development process tests, catching problems earlier and reduce both the cost of fixing problems and the risk of sliding delivery dates

The implementation of several simple tests you learn a lot about how to go to the most complex, large-scale tests as well:

• You will better understand what kind of traffic puts stress on your rear end and what does not work

• You will see how your rear end and react know where to look, what functionality registration to use, and generally how to get the most out of your tests

This means that when it comes time to perform a complex test, large scale, you will be better prepared to do. The large load test high-risk venture will become weaker and smaller than ever.

Do not let perfect be the enemy of the good

Realism is a good goal, but do not lose sight of the compromise. If you expect the test until you have a perfect simulation, you miss opportunities to solve problems. Look for ways to start with simple tests instead, then build to more complex simulations.

Keep in mind that early screening is no panacea, either. Problems can creep in late in the development and continuous testing can avoid last minute corrections.

Finally, remember that you have limited time and a demanding schedule. If you can get 80 percent of the results with 20 percent of the work, you will have more time to address other pressing problems.

0 views0 comments

Recent Posts

See All

Comments


bottom of page