This thesis examines methods for scalable testing of automated configuration of network devices. By applying modern DevOps practices to the domain of network devices, a new approach to network automation testing emerges. While containerization is well established in software development, its use for automating the testing of network configurations remains limited.
Traditionally, testing relies on physical devices, such as real switches and routers, to validate automated network configurations. An alternative approach is to use virtual machines, to which the configuration is sent and tested manually. This thesis proposes a container-based methodology for network configuration testing. This ensures that the testing environment remains identical across all test cases. Furthermore, errors can be reproduced easily, as the environment is described deterministically in code.
This approach provides repeatable test results, significantly reducing the time required by experts to perform manual configuration testing. In addition, it enables early detection of errors throughout the development process.
|