Home // International Journal On Advances in Internet Technology, volume 5, numbers 3 and 4, 2012 // View article

Performance Test Case Generation for Java and WSDL-based Web Services from MARTE

Authors:
Antonio García-Domínguez
Inmaculada Medina-Bulo
Mariano Marcos-Bárcena

Keywords: software performance; Web Services; MARTE; model driven engineering; test generation

Abstract:
Obtaining the expected performance from a workflow would be easier if every task included its own specifications. However, normally only global performance requirements are provided, forcing designers to infer individual requirements by hand. Previous work presented two algoritms that automatically inferred local performance constraints in Unified Modelling Language activity diagrams annotated with the Modelling and Analysis of Real-Time and Embedded Systems profile. This work presents an approach to use these annotations to generate performance test cases for multiple technologies, linking a performance model and a design model with a weaving model in a meet-in-the-middle approach so users can write their software according to their needs. Two implementations of the approach are described, which have been published as open source software. The first implementation extracts models from Java unit tests using MoDisco and weaves them with the performance models in order to convert some or all of the test cases in the selected JUnit test suites to performance tests. The second implementation extracts models from WSDL documents including message templates and template variables and uses these models for generating test inputs, test code and test infrastructure. Users can customise the service catalogues, test inputs and message templates to obtain a high degree of flexibility, and the test infrastructure provides automated configuration and test reports. These two successful implementations for different technologies validate the proposed approach as a generic framework for generating performance test case artefacts from existing software.

Pages: 173 to 185

Copyright: Copyright (c) to authors, 2012. Used with permission.

Publication date: December 31, 2012

Published in: journal

ISSN: 1942-2652