Definition
Integration testing where the interfaces between system components are tested.
[Source: BCS]
Comment
Experience shows that this is an area where it is very easy to find bugs.
Traditionally the incorrect mapping of data between the systems causes these bugs and these may result in the following types of bug:
-
Data is inconsistent between systems due to truncation or misinterpretation of the information.
-
The software that interfaces between the two systems fails - and no data is transferred - (this often results in the entire interface failing).
Normally, this is done in two phases:
-
When the interfaces are tested individually during system testing (essentially using a “dummy” system or stub to mimic the distant system or a closed-loop system) and
-
when the two systems are tested together with the systems communicating with one another during integration testing.
On many programmes and large projects, testing will take up a significant portion of the budget. But it will receive less management attention than an activity of this size warrants.