Research work in this project deals with the automatic generation of a TTCN-3 compliant test system intended to validate functional and performance requirements of a given system under test. Moreover, our testing methodology that is based on a Markov Chain usage model also provides additional statistical metrics such as the estimated reliability. The complete process relies on the UML 2.0 and the SPT-Profile which is a real-time extension for the UML.
Starting with a use case specification and scenario descriptions a protocol state machine is derived as the basic test model. By adding an operational profile the resulting model supports statistical analysis (e.g. the average test case length) as well as the statistical generation of test cases. Given some supplementary information, i.e. the data types and test data, the complete test suite can be generated in TTCN-3, a standardized testing language. After executing the test suite, automatic analysis of the test verdicts will provide major metrics such as the reliability estimation for the system under test.
For automation and for convenient application, all steps are implemented in XSLT stylesheets or in Java. The tool chain is based on the standardized interchange format for the UML called XML Metadata Interchange.
- Schlüsselwörter: software testing; automatic test generation; Markov chain usage model; UML2.0; TTCN-3; QoS
- Projektdauer: 2002-07-01 - 2008-06-30
- Dr.-Ing. Winfried Dulz
- Dipl.-Inf. Matthias Beyer
- Fraunhofer Institute for Integrated Circuits IIS
- "Performance Issues in Statistical Testing," Proc.of 13th GI/ITG Conference on Measurement, Modeling, and Evaluation of Computer and Communication Systems, Berlin, Erlangen, Germany, pp. 191-207, März 2006 ,
- "Scenario-based Statistical Testing of Quality of Service Requirements," Scenarios: Models, Transformations and Tools, Heidelberg, Schloss Dagstuhl, pp. 152-173, September 2003 ,