Comparative study of an automated testing coverage for a TCP/IP stack implementation
|Title||Comparative study of an automated testing coverage for a TCP/IP stack implementation|
|Summary||The topic of the project is the comparative study of the coverage of the tests generated by the QuickCheck tool against real coverage requirements|
|Keywords||model-based testing, coverage, TCP/IP|
|TimeFrame||Dec 2020 - May 2021|
|References|| Wojciech Mostowski, Thomas Arts, and John Hughes. Modelling of Autosar Libraries for Large Scale Testing. Proceedings, 2nd Workshop on Models for Formal Analysis of Real Systems (MARS 2017), Uppsala, Sweden, April 2017, Volume 244 of EPTCS. http://ceres.hh.se/mediawiki/images/b/bb/Mostowski_mars2017.pdf
Thomas Arts and John Hughes (2016): How Well are Your Requirements Tested? In: 2016 IEEE International Conference on Software Testing, Verification and Validation, pp. 244–254, doi:10.1109/ICST.2016.23.
The topic of the project is a comparative study of the coverage of the tests generated by the QuickCheck tool against real coverage requirements. The QuickCheck tool generates tests automatically based on a model - a high-level formal description of the system. The generated tests, even though easily provided in arbitrary quantity, may or may not reflect a typical real use of the component under test in terms of invoked operations and their inputs. For the case study we choose an available Erlang TCP/IP stack implementation and its QuickCheck model used for test generation. The real use of the stack can be easily analysed using a network analysis tool like Wireshark. The first challenge is to compare this to the the coverage of QuickCheck tests, the second challenge is to propose a way to guide QuickCheck test generation to improve this coverage (if necessary) towards more realistic one.