SCI-Aware Application Placement in the Cloud-Edge Continuum

The increasing environmental impact of ICT systems calls for carbon-aware management strategies for distributed multi-service applications. In cloud-edge environments, service placement decisions can significantly affect carbon emissions, yet existing approaches rarely rely on standardised carbon metrics. In this article, we address the carbon-aware placement problem by leveraging the Software Carbon Intensity (SCI) metric, recently recognised as an ISO standard. We propose a declarative model for microservice-based application placement that accounts for application topology, request patterns, internal workflows, and both operational and embodied emissions. Building on this model, we introduce an optimal exhaustive strategy and a configurable greedy heuristic solution to determine eligible and SCI-aware placements, differently balancing among carbon emissions and decision-making time. Our experimental assessment shows that the heuristic achieves optimal or near-optimal placements with significantly lower execution times than the optimal exhaustive search.