A systematic test case generation approach for testing message length variability

Variable length messages have been in use for a long time for efficient delivery of information. As there are many different ways, with varying complexity, to utilize message length variability, it is crucial to thoroughly test the capability of the parsers of such messages to ensure that they correctly handle the variability. However, testing techniques for message length variability were developed only in fragments in the past and therefore test developers who are faced with the task of testing variable length messages are left with little guidance or few techniques for handling them systematically. This paper develops an approach for systematic test cases generation for testing the parsers of variable length messages. To do so, we develop taxonomy of message length variabilities, and derive test requirement patterns for them using the taxonomy. Then test requirements for particular protocols can be derived from the patterns. A case study is conducted to show that the proposed taxonomy and the test requirement patterns are effective in deriving test cases for actual protocols. The results showed significant improvement over the conventional approach. It revealed many missed requirements that were not identified with the test requirements developed using the conventional intuition-based approach.