Traditional hosting providers use physical servers or virtual machines to host services. Each server manages multiple services, which means that all data is stored on the same medium. This is not a suitable approach if the services handle sensitive and confidential data such as bank card numbers and similar data. Such services need to be isolated on a private server, which can lead to inefficient use of system resources, higher maintenance costs, non-standardized and complex service deployment processes.
In this thesis we developed a prototype system for a container hosting provider based on Kubernetes, Docker and web technologies. The system is modular, communicates with standard Docker and Kubernetes components and is consequently less complex to use for the end user. With a simple and intuitive user interface, container deployment is straightforward. With Docker there is a standardised service deployment process and with Kubernetes there is scalability, high availability, fault tolerance and container isolation to ensure protection of classified and sensitive data and to reduce maintenance costs.
|
