EcoKube: Simulating Carbon-Aware Scheduling Policies in Heterogeneous Edge-Cloud Environments

EcoKube addresses the problem of rising carbon emissions from cloud and edge computing, particularly from AI workloads, by enabling reproducible evaluation of carbon-aware scheduling policies. The method introduces a configurable simulation framework with an event-driven deterministic simulator, policy hooks, and a heterogeneity-aware reference policy for heterogeneous edge-cloud topologies. Experimental evidence shows EcoKube's reference policy compared against the default Kubernetes scheduler, KEIDS, and TOPSIS/KCSS using synthetic batch workloads. This matters because it provides a consistent, reproducible workflow for evaluating sustainability-aware scheduling policies before real-world deployment, which existing solutions lack.