EarthCast and Penguin: Transforming HPC Operational Environmental Predictions Around the Globe
Moving EarthCast Technologies from Idea to Production EarthCast Technologies produces the most advanced realtime global satellite-based environmental prediction system in the world. However, moving from idea to actually creating unique, high-resolution global weather predictions for the aviation industry, agriculture, maritime, hydrology, and other industries required meeting both technological and business challenges. Penguin Computing partnered with EarthCast, using our powerful, bare metal high performance computing (HPC) cloud solution Penguin Computing® On-Demand™ (POD™), HPC services, highly-rated support, and our Linux and weather industry expertise, to tackle those challenges.
Until recently, only very large organizations (mostly governmental) could afford the computing resources and required HPC expertise to support complex forecasting capabilities that combined complex real-time data feeds of global observations with numerical modeling prediction technology.
Creating precise monitoring and prediction products outlining local hazardous weather and environmental impacts around the globe was simply out of reach.
Dr. Greg Wilson, Earth system scientist and president and founder of EarthCast Technologies, knew that the potential market for his products was limited by the enormous costs of computing resources required to create such complex models.
Producing timely forecasts for a larger market would require adding infrastructure priced in the low seven figures, managed by a dedicated support staff. However, this infrastructure and headcount would only be used 20% of the time, leaving a staggering – and unprofitable – 80% downtime.
EarthCast attempted to run their forecasts on traditional cloud providers but, like many others, realized that those resources would not support the tightly coupled code necessary to produce their products. Eventually, EarthCast found Penguin Computing and realized that Penguin’s HPC Cloud POD platform fulfilled their robust technical requirements but also gave them much faster, much more cost-effective access to the required infrastructure.
With traditional HPC clusters for compute, high speed, low latency interconnects, and true on-demand pricing, they could begin offering their products to commercial clients and expand their overall market and, thus, their business…