Energy-aware application life-cycle management in cloud edge computing continuum using applications profiles

Efficient management of an application's life-cycle on a Cloud Edge Computing Continuum (CECC) infrastructure requires knowledge about: (i) the application behavior as well as its requirements in terms of compute and network resources evolution over time; (ii) information about the state and characteristics of the infrastructure, including its produced carbon footprint. In this paper, we propose an architecture for the CECC Application Orchestrator to efficiently handle the application's life cycle by designing and relying on applications and infrastructure profiling, which is done using monitoring information, including energy metrics and carbon intensity of infrastructures. First, we define a modeling method for application profiles, which represent the application's current and future compute and network requirements. Second, we model the problem of selecting the best locations to deploy or migrate applications while minimizing the deployment's cost and the carbon footprint. The model decides the current and future placements of each application and its allocated resources based on the application profile, the availability constraints of the application, and the available infrastructure resources. Last, we propose a heuristic solution to solve the problem rapidly, and we compare the different trade-offs between carbon and cost efficiency.