Comparison of Docker Swarm and Kubernetes on the example of setting up a mail server

I have been working with self-hosting various services in containers for quite some time. This thesis seemed like an excellent opportunity to deepen my knowledge and to set up a highly available distributed system for the first time.

We use services running on highly available systems every day, though we are likely not aware of it until something suddenly stops working. The research and implementation of such a system allowed me to familiarize myself with many challenges and solutions that are essential for ensuring uninterrupted service operation.

Practically all of us use email nowadays, and most of us use free solutions from large companies like Microsoft Outlook, Google Gmail, or Proton Mail.

Email is one of the few services I had never hosted myself, and I had also never worked with cluster management systems, so I decided to study and compare them—I wanted to thoroughly examine the advantages, disadvantages, differences, and similarities between the two.

Both are very popular container orchestration systems. Kubernetes is more commonly found in production environments, so I was curious why Docker Swarm, which at first glance seems simpler to configure, is not as widely used.

In the theoretical part, I researched and described containers, the Kubernetes and Docker systems, and, since the subject of comparison is a mail server, I also wrote about email and email solutions. I touched on email spam which is something every email user has experience with.

For the needs of the BTEC hypothesis, I also delved into big data, where I explored what big data is, what its connection to email is, and how big data can positively or negatively impact our online security and privacy. Since this is a relatively complex project, in the practical part I wanted to present all the individual components used in great detail.

First, I introduced all the services I deployed in both systems, and I provided the reasons for my decision to use them. I described the product setup processes in detail, from setting up the clusters themselves to setting up the individual services, and the additional specifics for both systems.

Later, I also described the complexity of configuring the mail server, with an emphasis on security configuration. Toward the end, I tested the product in both systems with my test script for mail servers, conducted an analysis of the results, and wrote an in-depth comparison of the setup using both systems.