An Investigation of Testing Environments for Backend/Cloud-Services Increasing Efficiency of Testing in the Automotive Industry

Abstract

Products using cloud services are often problematic to prove correctness and quality of. In automotive industries specifically, the clouds are commonly developed alongside the products and controlled by separate companies. If the clouds are inaccessible or uncontrollable, efficient testing using both positive and negative testing is impossible. This Master’s Thesis strives, in cooperation with Delphi Automotive, to improve the testing efficiency in the automotive industry by designing a testing environment for communication from an application to back end/cloud-services. To resolve the issue with non-thorough testing, a cloud simulation environment was to be developed. Furthermore, to enable portability to different projects, the simulation should have complete configurability. The result was a general testing environment that can be applied to almost any product using HTTP/HTTPS communication. An evaluation of efficiency of the new testing procedure was performed and the efficiency was found to increase due to the overall higher thoroughness of the testing. However, the time to write tests and configure the environment was found to increase, as a result of the open and general nature of the testing environment. Worth mentioning is that this setup time is predicted to decrease in a real-world project. The testing environment created a time consumption overhead of 1-2 hours. By writing targeted test cases, several previously undiscovered bugs were found relating to retry-handling. Overall, spending some time developing a general testing environment seems to be profitable. Furthermore, extending the project with an easier configuration interface might be worthwile to further mitigate the setup time.