鲁棒性试验来源

Robustness Tests
Robustness Testing can be defined as the degree to which a software system or component can function
correctly in the presence of invalid inputs or stressful conditions.
Some key items to consider when developing robustness tests include:
- Invalid equivalence classes for variables (i.e., invalid data, stale data, data miscompare, data
drop-out, asynchronous/aperiodic data characteristics, test mode data, initial conditions,
abnormal data rate, transport delay effects, floating point values, NaN, + infinity, etc)
- Arithmetic overflow protection mechanisms
- Floating Point exception handling
- Out-of-Loop count values when possible
- Protection mechanism for exceeded time frames (e.g., timing & blocking problems, Time Budget
Exceeded (TBE))
- Provoking of invalid software transitions when possible
- Transitions back to normal conditions
- Transition back from a detect to non-detect for a fault
- Worst-case timing scenario with and without failure conditions; these failure conditions should not
affect the worst-case timing scenario.
Neither list above is necessarily an exhaustive list of items to consider when writing normal or robustness
test cases. Finally, some of the items described above for the robustness test cases are more suitably
performed in a Hardware/Software Integration test environment.
Determining whether test cases satisfy the objectives of normal and robustness test cases is done during
the inspection of the software test procedures.