Defining an using an ontology of test environments

Abstract

In the course of this master thesis, an ontology of test environments within the scope of automated software testing in the automotive industry was created. Based on an analysis of the current test environments and on a literature research, the focus was placed on a specific use case, which was defined to provide decision support for testers. When manually testing, the test personnel has to have knowledge about hardware, software and test systems to be able to perform tests on the correct environment. Since there are a lot of interdependencies between the necessary entities of test environments, these dependencies were analysed. The knowledge was gathered and concentrated in the form of an OWL ontology. The focus of the use case of this ontology lied in mapping a specific test case and software under test to a specific test environment. To accomplish this, mechanisms of semantic reasoning were investigated and used. Furthermore, the dependency specification within the ontology and the import of data from different sources has been analysed. The ontology was modelled using Protégé, whereas for the import and usage of the ontology the knowledge graph application Stardog was used. The advantages of Stardog are the integrated HTTP API and the possibility for storing data from heterogeneous sources in a unified way. Further, it allows querying and manipulation of the knowledge graph using SPARQL and supports built-in reasoning over ontologies. To fulfil the proposed use case, the imported ontology in Stardog was queried for relevant information. These queries were prepared and evaluated with the test personnel and stored on the server. A separate tool for test automation was used to call the queries and output their results. These results provide decision support for testers through the specified dependencies in the ontology. Through the usage of the tool querying the ontology, an improvement of the test automation process is expected.