Self-Healing of Distributed Systems

Abstract

Self-healing is a buzzword in distributed computing domain since last decade. During this study, the evolution of distributed systems with respect to self-healing was studied. Some of the common architectures used for self-healing of distributed sytems is presented. With the inception of cloud computing, cloud computing platforms have vastly adopted self-healing methodologies when running their large scale server farms. The study was inspired by the project which the author has been working at his company. In the upcoming sections author has proposed some automated strategies to heal a distributed system replacing the manual process which involves human intervention. For the implementation of the suggested methodology, author has proposed a solution based on docker and kubernetes. Self-healing solution has been implemented on a WSo2 EI cluster. The research has been able to achieve the self-healing of a VM cluster in VM level. Going further from initial objectives, auto scaling has been achieved in the VM cluster as well.