Nowadays, container technologies are getting more and more popular. Their efficiency and the easy portability and installation of the services they host has made the use of containers unarguably attractive. Using container technologies, one can build enormous clusters of several hundreds of physical servers that are capable of hosting thousands of containers. Maintaining and updating the clusters poses a difficult task for the administrators. In the world of containers, Docker meant a huge breakthrough. Soon, the need arose to have a management platform with which it would be possible to manage containers, even in infrastructures that span multiple servers. In my thesis, I demonstrate Kubernetes, which is a popular and widespread platform that fulfills said need. My task included designing an automated process for the installation of Kubernetes clusters. Therefore, the automation of the various phases is discussed in detail. Using a network implementation (backend) is essential for the operation of Kubernetes. These implementations often rely on network overlays. For this reason, I also provide an overview of network overlays before moving on to discussing the chosen network implementations. In addition to describing the network implementations, I perform performance tests with a variety of test cases. Based on the results, I conclude that it is not possible to declare an overall winner since each implementation has areas of application where it falls short of others. Thus, it might be worth employing the solution CNI-Genie which makes it possible to use multiple network implementations at the same time. This work shall be a good starting point for those people who already have experience in administration, but are only just starting to work with container technologies. In my thesis, I aim to pass on the basics to the reader, in order to make it easier to get acquainted with one of today’s IT world’s most hyped topics.