Sustainable cloud infrastructure: AI-driven carbon-aware kubernetes scheduling and resource management

Sakhamuri, Naga Sai Bandhavi (2025) Sustainable cloud infrastructure: AI-driven carbon-aware kubernetes scheduling and resource management. World Journal of Advanced Research and Reviews, 26 (2). pp. 2138-2145. ISSN 2581-9615

This technical article explores an innovative framework for reducing carbon footprints in cloud infrastructure through AI-driven, carbon-aware scheduling and resource management in Kubernetes environments. As cloud computing continues its exponential growth, the environmental consequences have become increasingly significant, with data centers consuming a substantial portion of global electricity. The intersection of cloud infrastructure, artificial intelligence, and environmental sustainability creates both challenges and opportunities. The article examines current energy consumption patterns in data centers, carbon footprint considerations related to different energy sources, and regulatory pressures driving sustainability initiatives. It highlights the limitations of traditional Kubernetes resource management, which prioritizes performance metrics while neglecting environmental impact. The proposed carbon-aware framework leverages machine learning to optimize workload placement based on environmental factors, introducing predictive energy consumption modeling, temporal workload shifting, and carbon-aware autoscaling. Implementation strategies and real-world impacts are discussed, including phased deployment approaches, quantifiable carbon reductions, and cost savings through more efficient resource utilization, demonstrating that environmental responsibility and operational efficiency can be simultaneously achieved in modern cloud infrastructure.