Software Integration Testing: Complex but Essential

In integration testing, input modules that have been unit tested are grouped into larger module. Tests like performance testing can be done at any stage of the development. However, integration testing has to follow a certain schedule in order to be effective. It is carried out after unit testing but before system testing. Tests as defined in an integration test plan are applied to the integrated module and deliver an integrated module for system testing.

Any large system consists of smaller sub-systems and the subsystem consists of components. Initially the components are tested independently and made bug free. Then this process is repeated for a group of components or subsystem and then to the integrated system.

The objective of Integration testing is to make sure that the interaction of two or more components produces results that satisfy functional requirement. The defects in the interface of the components and also their interaction will be observed during the integration testing. The integration testing plan comes from the physical makeup of the system components and these were identified and designed earlier in the development stage.

The task of Integration testing is centered on the system functions. These functions are individual program modules or individual units that when assembled together make a final integrated software. These units have been tested in the unit testing stage and all bugs in individual unit software should have been removed at that stage.

These units should perform satisfactorily as individual entities. However, there is no guarantee that the integrated software will function satisfactorily since it has passed through the unit stage. Any conflict between various modules has to be resolved at the integrated testing stage.

The purpose of integration testing is to verify functional, performance and reliability requirements placed on major design modules. These modules are integrated or assembled from a group of units as a major module. Test cases are constructed to test that all components within assemblages or major design modules interact correctly. The whole concept is similar to a "building block" approach, in which verified assemblages are added to a verified base which is then used to support the integration testing of further assemblages.

The types of Integration testing are big bang, bottom-up, top-down and hybrid. Brief details on this are as given below.

	Big bang - In this type of testing, all/ most of the developed modules are integrated to form a complete/ major software system as the case may be and then used for integration testing. The Big Bang method saves a lot of time in the integration testing process and also cost effective. However, this process warrants proper recording of test cases and also their results in the absence of which the process of integration becomes very difficult.
	Bottoms-up – In this the testing is carried out starting from the lowest level software modules or units of the system. Once these are tested and debugged they are integrated together and tested as an integrated unit. A number of such integrated units after testing are integrated together to form the higher level integrated module and again tested for errors. This process is continued until the whole application is tested and debugged. For this approach to be effective all units at that level have to be ready before testing.
	Top-down – This is exactly opposite to bottoms-up approach. Top down integration testing is an incremental integration testing technique which begins by testing the top level module and progressively adds in lower level module one by one.
	Hybrid – Both Top-down and bottom-up approaches have their own advantages and disadvantages. In top-down integration testing it is very easy to follow the top-down software development process whereas in bottom-up testing the code used is tested repetitively. The hybrid integration testing takes advantage of the benefits of both top-down and bottom-up testing.