A Network Science-Based Approach for an Optimal Microservice Governance
With the introduction of microservice architecture for the development of software applications, a new breed of tools, platforms, and development technologies emerged that enabled developers and system administrators to monitor, orchestrate and deploy their containerized microservice applications more effectively and efficiently. Among these vast arrays of technologies, Kubernetes has become one such prominent technology widely popular due to its ability to deploy and orchestrate containerized microservices. Nevertheless, a common issue faced in such orchestration technologies is the employment of vast arrays of disjoint monitoring solutions that fail to portray a holistic perspective on the state of microservice deployments, which in turn, inhibit the creation of more optimized deployment policies. In response to this issue, this publication proposes the use of a network science-based approach to the creation of a microservice governance model that incorporates the use of dependency analysis, load prediction, centrality analysis, and resilience evaluation to effectively construct a more holistic perspective on a given microservice deployment. Furthermore, through analysis of the factors mentioned above, the research conducted, then proceeds to create an optimized deployment strategy for the deployment with the aid of a developed optimization algorithm. Analysis of results revealed the developed governance model aided through the utilization of the developed optimization algorithm proposed in this publication, proved to be quite effective in the generation of optimized microservice deployment policies.
