Towards energy-aware execution and offloading of serverless functions

The convergence of Function-as-a-Service (FaaS) with Edge computing presents a promising avenue to address the demands of geo-distributed and pervasive applications. Yet, leveraging both the reduced latency of Edge and the scalability of FaaS necessitates novel architectures and implementations, given the higher heterogeneity and resource constraints of Edge environments. An important issue regards energy management, as serverless functions may be executed in energy-constrained computing nodes (e.g., powered by batteries or renewable yet uncertain energy supplies).In this paper, we present a solution for energy-aware function execution relying on two key mechanisms: (i) the ability of choosing among multiple function implementations at run-time, trading off energy savings with approximate computation results; (ii) computational offloading, to move function execution from the Edge to the Cloud, based on energy-aware policies. We integrate our solution in Serverledge, an open-source FaaS platform designed for the Edge-to-Cloud continuum. Our evaluation shows that energy-aware offloading and execution are effective in extending the lifespan and throughput of battery-powered nodes.