This analysis demonstrates carbon-aware workload placement can lower CO2 emissions by 10-30% in Kubernetes, highlighting the importance of renewable energy integration.
Key Points
Carbon-aware scheduling can reduce CO2 emissions by up to 30% while maintaining performance levels.
Workloads are dynamically scheduled based on predictive carbon intensity values from sources like ElectricityMap.
The framework utilizes Kubernetes mechanisms like node affinity and custom scheduling policies for optimal placement.
Challenges include ensuring data granularity, interoperability standards, and enterprise adoption of carbon-sensitive methods.
Cite This Study
Nishanth Reddy Pinnapareddy (2025) studied this question.