ENERGY-EFFICIENT CONTAINER ORCHESTRATION FOR GREEN CLOUD COMPUTING USING KUBERNETES

Abstract

The rapid expansion of cloud computing, containerized microservices, distributed digital platforms, artificial intelligence workloads, enterprise applications, and data-intensive services has significantly increased computational resource demand and associated energy consumption in modern data centers. Kubernetes has emerged as a dominant container orchestration platform because it provides automated workload deployment, scheduling, scaling, service recovery, resource management, and application portability across heterogeneous cloud infrastructures. However, conventional orchestration strategies primarily emphasize workload feasibility, availability, performance, and resource balancing, while energy consumption, carbon intensity, idle-node power, workload flexibility, and sustainability objectives may receive insufficient consideration. This paper proposes an energyefficient container orchestration framework for green cloud computing using Kubernetes. The proposed methodology integrates workload profiling, real-time infrastructure telemetry, node energy characterization, carbon-awareness, predictive demand analysis, intelligent pod placement, adaptive workload consolidation, energy-aware autoscaling, sustainable GitOps governance, and continuous feedback. Containerized workloads are classified according to computational demand, latency sensitivity, business criticality, execution flexibility, and scaling behavior. Kubernetes worker nodes are profiled according to available resources, current utilization, energy efficiency, operational state, and carbon-related context. The intelligent orchestration mechanism evaluates candidate placements through multiple operational indicators and selects nodes that reduce unnecessary energy expenditure while preserving application performance and service reliability. The framework further introduces adaptive consolidation to reduce lightly utilized active nodes and predictive scaling to anticipate recurring workload changes. A representative experimental evaluation demonstrates that the proposed approach can reduce total energy consumption, improve effective resource utilization, decrease average active-node count, reduce carbon-impact indicators, and maintain acceptable response time and service availability compared with conventional Kubernetes scheduling. The findings establish that energyaware orchestration can transform Kubernetes from a general container management platform into an important technological foundation for sustainable and environmentally responsible cloud computing.