Microservice Architecture for Stateful Applications

Abstract

The widespread adoption of microservices in modern software development has required efficient solutions for managing databases within Kubernetes environments. While Kubernetes offers some mechanisms for handling stateful applications, challenges exist in performance and reliability, especially for time-critical workloads such as Geographic Information System (GIS) data processing. This study investigates the limitations of existing approaches and a solution leveraging Ceph as a backend storage system integrated with a Kubernetes cluster. Through the creation of an External Ceph Storage Cluster and reconfigured Helm charts, the proposed solution enhances data management for reliability, scalability, and performance. Results demonstrate improvements in data processing speed and fault tolerance, contributing to the evolution of Kubernetes-based databases. Additionally, a deployment script is provided to facilitate easy experimentation and adoption of the solution, expanding exploration and integration of advanced storage technologies within Kubernetes ecosystems.