Skip to main content

Performance Testing 101

So one of my objectives is to be able to offer more in regards to Performance Testing.

Whilst I understand the basics, I figured I'd write them down and start a series of blog posts hopefully detailing my findings and explaining why performance testing is important, and the different types of performance testing.

In order to understand why performance testing is important, it will be good to know some of the more popular types of performance testing are....

Load Testing
This is testing a system to understand and analyse the behaviour of a system under a specific expected load. This is probably the most common form of performance testing, it is useful in identifying any bottlenecks in a piece of software. 

Stress Testing
This is testing a system to breaking point, so that you can understand the point at which the system fails and under what load.

Endurance Testing
Similar to load testing, but this will be load under a prolonged period of time, in order to gain confidence that the system can perform under loads for a long period and measuring the performance of the system.

Spike Testing
This is achieved by a sudden increase in load against a system, to make sure that it can handle the sudden increase in users, an example would be of a new system going into production that suddenly has to deal with a huge influx of users.

When might you use one of the above?
Now you have a basic understanding of some of the different types of performance testing, we can now look at where they may be useful, and when you might consider doing one of the above.

Load Testing is possibly one of the more well known and what often gets confused as Performance Testing full stop, when someone mentions performance testing, it is often assumed that they are talking about load testing, however, unless we dig deeper, assuming something like that can make an ass out of you and me...

Anyway, I digress, Load Testing would be useful when there is a change to the system that can cause there to be an increased amount of load, for instance we recently changed data types from VARCHAR to NVARCHAR, this meant there was double the amount of data being stored in databases and being used in any database call, so we decided its important to do some forms of load testing on the front end, as well as the back end on the SPROCs themselves.

Stress Testing is useful to identify what your system is able to cope with, for example, if you were expecting X number of visitors during a certain period of time, if you know what your break point is, with regards to stress testing, you can safely say that the system can perform under the expected load and then some.

Endurance Testing is useful in identifying memory leaks in a system, if you expect the system to be under heavy loads for a prolonged period of time, it is essential to perform endurance testing to identify any problems that may arise.

Spike Testing as mentioned above is useful when putting a new service live, as you expect a sudden increase in traffic once it goes into production and will be hit with a sudden influx of load.

So, there you have it, some basics about the different types of performance testing, and why it might be important in your project.... Going forward we'll investigate different toolsets for performing the above and when is the best time to do performance testing in an Agile project.

Comments

Popular posts from this blog

Treating Test Code as Production Code

It's important when writing automated tests to remember that the code you write should be up to production standards, meaning any conventions that you have in place should be adhered to and that it should follow good design patterns. Too many people often say why does it have to be as good as production code, it's "Only" a test, so long as it passes then that's fine... To answer this we need to look at why we want our tests to be written in such a structured and efficient manner: - Maintainability - by making the test code structured and efficient, it becomes far easier to maintain and in doing so changes in the future can happen quickly as the test isn't linked to anything that it shouldn't be and it's easy to understand for a new set of eyes. - Durability - Again by making the tests structured they should be resistant to changes, if you change a variable name for instance then it shouldn't effect the unit test unless it absolutely has to....

Testers: Be more like a Super-Villain!

Who doesn't love a Super Hero? Talk to my son, and he'll tell you how much he loves them, talk to many adults and they'll say the same! Deep down, we all love to be the Super Hero, we all want to save the day! However, I want to talk about the flip side of Super Heroes, the Super Villains... I often play Imaginext with my son, and I (unfortunately?) am nearly always the Super Villain! Be it Lex Luthor, Joker, Two Face, Mr Freeze or The Riddler! These are all great characters and great Super Villains, but why would I want to write about Super Villains? A while ago where I worked, we had a few Super Heroes, people who would be able to come in and "fix" things that had broken and help deliver projects on time. We then shifted, we decided to do away with the Super Hero culture and try and prevent from being in that position in the first place, whilst we didn't go as far as wanting to hire Super Villains, it's definitely a story that has stuck with me and t...

Using BDD and gherkinising your Acceptance Tests

In my post Testing of Automated tests , I mention about a BDD framework which involves using BDD to drive your acceptance tests. BDD stands for Behaviour Driven Development.  One effective method of writing BDD tests are by using a format known as Gherkin language. These consist of Given, When, Thens. The main advantage of the gherkin language is that it's readable by the business, and in an ideal world forms part of the Conditions of Acceptance around a PBI. Also, using a Visual Studio plugin of SpecFlow , you can integrate your Gherkinised COAs into your solution with feature files, and then drive the automated tests, however, for this post I will focus solely on how to effectively gherkinise your acceptance tests. A Feature file consists of a feature outline, which details what the feature file is testing followed by Scenarios and examples (parameters).  The BDD scenarios are made up of a Given, When, Then... These are effectively an initial state (Given), an action (W...