Estimating runtime, cost for Weather Research and Forecast tasks on AWS

Deepika Udupa; Gregoris Malekos; Kostas Flokos

doi:https://doi.org/10.63438/YZFU3809

Estimating runtime, cost for Weather Research and Forecast tasks on AWS

Deepika Udupa, Gregoris Malekos, Kostas Flokos 4 pages / April 4, 2025 https://doi.org/10.63438/YZFU3809

Abstract

Cloud scheduling is an NP-hard problem. In this paper, we propose an end-to-end system that executes a WRF (Weather Research and Forecast) task by providing recommendations of resource configurations. The optimal executions scenarios containing 3 options: fastest, cheapest and a trade-off is generated by estimating the runtime using a data-driven ensemble regression method.