What are different characteristics of integration testing?

Sometimes abbreviated as "I&T", it is one of the most important testing in the world of software testing which tests units as a group. Means to say, in integration testing the units or modules are combined and tested. This is done to determine that the units are working in collaboration with each other or not.

- Integration testing is carried out after the unit level testing but before the validation testing.
- For integration testing units or modules that have been unit tested are supplied as input. Then they are conjoined to form groups and integration is performed over them and the output is given out.
- Integration testing is carried out to examine the functionalities, reliability, performance and requirements of those grouped modules under testing.
- Integration testing is implemented through the interfaces of the units with the help of black box testing techniques.
- During integration testing success and failures are overcome, inter process communication between the units is tested and using the input surface individual and distinct subsystems are implemented.
- Like other kinds of testing, integration testing also requires the aid of test cases to test various aspects of its input unit aggregates.
- Integration testing works on the idea of “building block” according to which verified and examined unit aggregates are implemented on a verified base which is the software system itself.

There are various kinds of integration testing techniques. Three major techniques have been discussed below:

- Big Bang:
In this type of integration testing all of the grouped units or modules are joined together to the form the complete and finished software system and then the integration testing is carried out for the whole software system in one go. The big bang technique is very effective when the software developer wants to save his/ her time. This technique is a true time saver package. But, it has a disadvantage which is that if the unit tests have not been recorded and carried out properly, the whole integration testing process may become more complicated and difficult to crack and may prove to be a hindrance between the integration testing and its goal.

- Big bang technique has a distinct testing method which is called “usage model testing”.
- This integration testing works for both hardware integration testing as well as for software integration testing.]
- This type of integration testing aims implement user like workloads in user like environments which are well integrated.
- This testing methodology involves the proofing of environment first and then later proofing the individual units by implementing their functionality or usage. - This can be thought of as an optimistic approach to integration testing.
- This technique requires more hard work as the problems are more. It is based upon the idea of “ isolate and test”.

But the big advantage here is that usage model testing provides good test coverage and thus making it more efficient than the other techniques.

- Top down integration testing:
This type of integration testing includes testing of the modules at the top most position first and coming down to the branches of those modules or units until the end of the program is reached.

- Bottom up integration testing:
This type of integration testing the modules or the units’ aggregates at the lowest level are tested first and then the upper modules are tested. This testing is carried on and on till the top most module is tested. This kind of testing is helpful only when all the modules have been developed and are ready for integration.

- Top sandwich integration testing:
This integration testing technique involves combination of top down and bottom up testing.