Advanced Computing in the Age of AI | Sunday, October 2, 2022

U.K. Weather Office Gets Exascale Boost 

Bernhard Staehli/Shutterstock

Improved weather forecasting and climate change analysis are among the emerging applications for the combination of exascale computing, data analytics and the storage demands each places on data management tasks.

European forecasters have been at the forefront of leveraging these technologies. The latest example comes from the U.K.'s national weather service, the Met Office, which announced this week it has selected a trio of HPC vendors to boost the capabilities of its Scientific Processing and Intensive Compute Environment, or SPICE initiative. Selected earlier this year were Bright Computing, DataDirect Networks (DDN) and SGI (NASDAQ: SGI), the office announced Wednesday (Oct. 19).

The exascale U.K. weather platform enables weather and climate researchers to reduce the time required to analyze massive amounts of climate simulation data. Among the requirements for improving the accuracy of weather forecasts and understanding the dynamics of climate change is the ability to use data-intensive, high-resolution models.

The Met Office said it would use the trio of exascale tools for further processing of daily meteorological data and "downstream" analysis of big data sets. The SGI platform is being used to manage multiple servers to increase overall computing horsepower supporting the SPICE initiative while serving as an upgrade to the Met Office's Managed Archive Storage System (MASS).

MASS also uses DDN's storage technology and serves as a repository and archive for climate research as well as worldwide observational data, the companies said.

The forecasting and climate research platform is tied together using Bright Computing's cluster manager for HPC applications as a way to deploy new SPICE clusters on bare metal servers. It also uses Linux cluster manager Bright System's OpenStack distribution for provisioning and managing the U.K. office's private cloud infrastructure.

By combining SGI's computing platform, DDN's storage and the Bright's OpenStack distribution, the Met Office was able to scale SPICE storage while handling simultaneous data reads/writes, the vendors said. The exascale configuration allows weather researchers to quickly spin up virtual machines or boost capacity by adding more servers to the OpenStack platform.

The combination is said it have boosted performance by crunching massive weather data sets in hours instead of days, the vendors claimed. U.K. weather researchers report they can now "perform multiple runs in one day" compared to previous post-processing of weather and climate data runs that took as much as two days to complete.

Along with handling massive weather data workloads, the new Met Office platform allows climate researchers and IT managers to operate the system simultaneously, the vendors noted.

The U.K. upgrade also underscores the growing gap between European and commercial weather modeling tools and those employed by U.S. forecasters. Critics of U.S. weather forecasting capabilities point to poor data assimilation used to integrate atmospheric information into models and an overall lack of processing power that hampers the ability to use high-resolution ensembles for delivering more accurate weather forecasts.

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).

Add a Comment