STAF-on-Eucalyptus: A Cloud-Based Software Testing Environment for Distributed Systems

Abstract: Cloud computing fundamentally provides access to computing resources and services over the internet. It offers potential solutions for effective testing by provisioning enormous amount of computing resources to end users. Utilizing the cloud for testing started around 2002 and has mainly focused on techniques for online testing, ranking, automated test case generation, monitoring, simulation, and policy data provenance (Yu et al, 2010). Commercial provisioning of testing services in the cloud today is mainly inclined towards providing a generic solution that attempts to address a wide range of testing purposes and therefore is not applicable in testing unique applications such as complex distributed systems, neither are they applicable in running unique test suites. Based on an industrial problem which is discussed on section 1.1, this study attempts to find appropriate strategies aimed at utilizing the cloud for testing unique applications or running unique test suites. In this paper, we discuss two approaches useful in solving this problem; the migration followed by end-to-end test automation approach. Finally, we propose a solution to this problem; STAF-on-Eucalyptus which is based on the Eucalyptus cloud computing system (Nurmi et al, 2009) and the Software testing automation framework (Cervantes, 2009).