Cloud Orchestration for Optimized Computing Efficiency: The Case of Wind Resource Modelling

The Weather Research and Forecasting (WRF) model is a multi-purpose open source weather model, which is widely used by the wind energy community. However, the state of the art applications that use the WRF model for wind resources do not always meet the sector's needs. The fundamental challenge is the massive quantity of data required for accurate modelling, as well as the expenses associated with the computing of such massive volumes of data, implying that state-of-the-art approaches are not appropriate. Our work tackles this issue with the use of a novel orchestration framework relying on cloud computing infrastructure. Cloud computing platforms provide robust infrastructure that allows for the deployment of production-level services as well as extended processing and storage capacity. This paper presents the overall software architecture of a Cloud Orchestration System, enabling the dynamic and flexible cloud deployment of containerized wind resource model chains. Moreover, a set of performance, functional and operational criteria are presented and evaluated, using a real use case of a wind resource model.