APIbuster Testing Framework

• Autor(a) principal: Sousa, Pedro Ferreira de
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/22512
• topic: APIbuster; Data modelling pipeline; REST API; SOA; Testing platform

In recent years, not only the Service-Oriented Architecture (SOA) became a popular paradigm for the development of distributed systems, but there has been significant progress in terms of their testing. Nonetheless, the multiple testing platforms available fail to fulfil the specific requirements of the Moodbuster platform from Instituto de Engenharia de Sistemas e Computadores, Tecnologia e Ciência (INESC TEC) – provide a systematic process to update the test knowledge, configure and test several Representational State Transfer (REST) Application Programming Interface (API) instances. Moreover, the solution should be implemented as another REST API. The goal is to design, implement and test a platform dedicated to the testing of REST API instances. This new testing platform should allow the addition of new instances to test, the configuration and execution of sets of dedicated tests, as well as, collect and store the results. Furthermore, it should support the updating of the testing knowledge with new test categories and properties on a needs basis. This dissertation describes the design, development and testing of APIbuster, a platform dedicated to the testing of REST API instances, such as Moodbuster. The approach relies on the creation and conversion of the test knowledge ontology into the persistent data model followed by the deployment of the platform (REST API and user dashboard) through a data modelling pipeline. The APIbuster prototype was thoroughly and successfully tested considering the functional, performance, load and usability dimensions. To validate the implementation, functional and performance tests were performed regarding each API call. To ascertain the scalability of the platform, the load tests focused on the most de manding functionality. Finally, a standard usability questionnaire was distributed among users to establish the usability score of the platform. The results show that the data modelling pipeline supports the creation and subsequent updating of the testing platform with new test attributes and classes. The pipeline not only converts the testing knowledge ontology into the corresponding persistent data model, but generates a fully operational testing platform instance