Introduction to Integration testing? What are approaches used in integration testing?

Unit testing is useful to test the modules individually. Problem arises when these modules are put together, which is also called interfacing. There is a huge chance that data is lost across an interface or one module can affect the functionality of the other module or global data structures can present problems.

Integration testing is a technique that is used for constructing software architecture while at the same time conducting tests to uncover errors that arise because of interfacing. There are two approaches to integration testing :

- Taking the big bang approach (all components are combined in advance and then testing is done) to integration is a lazy strategy that is bound to fail. One should integrate in an increment manner and do testing as you go. In big bang approach, cause of fail is not easily tracked and hence the complexity increases.

- Incremental approach can be started early and controlled easily. Incremental testing is usually better for large, complex systems. Two related modules are combined and tested and check whether they are working in the correct manner. Then, another module is added and checked.

Integration testing follows unit testing and precedes system testing. It helps in verifying functional, performance and reliability requirements. Reasons for integration testing includes:
- Fields are defined differently in different modules.
- Different perceptions and understanding of business requirements.
- Field content can have different assumptions.
- There are few errors that can be left uncovered during unit testing.