Elastic deployment of container clusters across geographically distributed cloud data centers for web applications

Containers such as Docker provide a lightweight virtualization technology. They have gained popularity in developing, deploying and managing applications across Cloud platforms. Container management orchestration platforms Kubernetes run application containers virtual clusters that abstract the overheads underlying infrastructures to simplify deployment of container solutions. These are well suited for modern web can give rise geographic fluctuations use based on location users. Such often require dynamic global A key issue is decide how adapt number placement maintain performance, whilst incurring minimum operating adaptation costs. Manual decisions naive to: over-provisioning hence cost issues; improper performance issues, and/or unnecessary relocations resulting issues. Elastic solutions essential support automated intelligent geographically distributed Clouds. In this article, we propose an approach continuously makes elastic plans aimed at optimizing even during processes, meet service level objectives (SLOs) lower Meta-heuristics used cluster adjustment. We conduct experiments Australia-wide National eResearch Collaboration Tools Resources Research using Kubernetes. Results show with only 0.5 ms sacrifice SLO 95th percentile response times able achieve up 44.44% improvement (reduction) compared approach.