Software Performance Testing: How Software Performs With Real Tasks

Performance testing comes under the categorization of non-functional testing and is usually conducted for web applications. There are different types of testing performed while conducting performance testing. One such test is load testing.  In this testing the system behavior is checked under load. The application is tested under heavy loads, such as testing of a web site to determine the point of degradation or failure of the response time. As the traffic increases the load on the website increases and you can determine the point at which the application cannot take the load. In classic performance testing, the time taken for responding to user command is measured. The number of transactions that can possibly be performed on the site at one time is also measured.

Another type of performance testing is stress testing wherein the system is stressed under heavy load like putting large number of data beyond storage capacity, complex database queries, continuous input to system or database load. The Testing is conducted to evaluate a system or component to determine the load under which it fails and how. The most preferable result is a graceful degradation under load leading to non-catastrophic failure. In case you operate on different programs simultaneously there will be certain slowdown on the program execution. Another is volume testing wherein the tester will add more and more data to find the point of non-functioning of the program. The volume test can also determine how much volume of software the system can hold before failure.

Performance tests normally are performed with the following objectives in mind.

	Identification of system bottlenecks
	Verification of current system capacity.
	Verification of scalability of the system.
	Determine optimal hardware/software configuration for your product.

Some of researches done by software testers indicate that the reasons for failure of performance tests are mainly due to hardware/ database problems, abnormal size and volume of the transactions which cannot be handled by the system and lack of proper bandwidth. Validation and verification of system qualities such as reliability and scalability are also performed in this testing. The incorporation of performance has to start even at the computer design stage i.e. well before any program coding is started.

Performance testing is normally conducted after the functional testing completion and is done during the System Testing phase. In addition to indicating whether the system meets the performance standard as envisaged originally it also can find out which of the two given systems are superior in performance or vice versa. It also can identify the part of the system which makes it better or worse. Introduction of performance testing in the early stages of development will go a long way in getting a better product at minimal cost. The cost of rectifying the performance defects is large if the testing is started in the late stages of development. Ideally, the test cases should be as similar to the actual expected use as far as possible, but this is very rare.