Green load balancing: Energy-efficient task distribution in cloud systems

With advancements like cloud computing expanding horizons for networking infrastructure, the associated energy consumption and carbon emissions of these data centers become more and more of a concern. Implementing carbon-aware load balancing strategies in cloud infrastructure allow for organizations utilizing these systems to take these costs into consideration without sacrificing performance or reliability. This study investigates the feasibility of incorporating carbon-aware load balancing within Google Cloud Platform (GCP) to minimize the environmental footprint of global traffic routing. By deploying web servers across three geographically distinct GCP regions–europe-west3, us-central1, and asia-southeast1 the study evaluates performance and environmental outcomes of simulating traffic under both traditional and sustainabilityprioritized configurations. Locust, a Python-based load testing tool, was used to record performance metrics like response time, request throughput, and failure rate across the three regions. Carbon-aware load balancing was implemented through manual capacity scalers which prioritized green-energy-powered regions like europe-west3. Results indicate that over 80% of requests were effectively routed to low-carbon zones, leading to a reduction of over 60% in carbon emissions and minimal impact on latency and reliability. The findings demonstrate that existing GCP tools can support green load balancing infrastructure and provide a promising direction for sustainable cloud-native architectures. This record contains metadata only. The publisher (IRJMETS) holds exclusive distribution rights. The final article will get an IRJMETS DOI; this Zenodo DOI provides a permanent reference in the meantime.