Integration Testing: Essence and Approaches

January 18 09:59 2017 by Nataliia Vasylyna Print This Article

Nowadays it is a good practice to test the system before the production code is written. This means that software testing is performed in parallel with writing code lines.

Usually, this technique presupposes several types of checking: unit or module testing is executed prior to a test team’s moving to integration testing. Integration checking includes testing of several units which are combined together in order to fulfill functions.

It is rather difficult test process which may be conducted via all three techniques: white, black and grey-box testing. The main goal of checking is to detect weak points or possible errors that may occur under various units combination.

Unlike unit testing, integration one may be fulfilled by both testers and developers. From this perspective, integration tests are performed on the basis of one of three approaches.

3 Approaches to the Integration Testing Application:

Top down – the most logical type of testing as it actually copies actions, occurred in a real environment. The modules are tested from top levels to the bottom ones. However, the domain functional is checked last.
Bottom up – the best way to check problems which may occur at the lower levels. Exactly this reveals the main minus of this approach; the top level errors may be discovered at the end. The DRIVERS programs are applied during this procedure.
Big Bang Integration – all parts of the system are checked simultaneously at once. It may seem the same as system testing but it does not. This approach has a probability of missing some serious failures as all units are tested at one time.