The aim of this Bachelor thesis is to describe the concepts, best practices, and tools needed to run full-scale applications in a containerized environment. The theoretical part of the thesis includes an overview of basic container environment concepts such as single process virtualization, networking between containers and the outside world, copy on write filesystem and basics of running stateful applications. Explanation of concepts specific to a multi-node cluster is also given. After the introduction of needed concepts, fundamental challenges of a containerized environment are presented together with proposed best practices to tackle them. Emphasis is put on privacy, security and reliability challenges specific to containers such as storing and managing persistent data across containers, kernel and network security, etc. Overview of alternative solutions is given along with use cases in which such tools would fit better than the included implementation.

The practical part of the thesis takes all the concepts from the theoretical part and shows their implementation in practice. This part describes the approach of an IT administrator of a small sized provider of managed IT services currently running standard physical or virtual servers on owned bare metal infrastructure making the switch to a containerized environment by deploying a centrally managed Docker cluster. The implementation includes scalability of deployed production applications, as well as redundancy, dealing with persistent data across containers and multiple physical servers as well as implements security best practices proposed in the theoretical part of the thesis.

The last chapter will outline areas for further research including technologies not yet well established in production environments as well as formal requirements for products that would better solve some of the challenges identified earlier.