The Met Office is combining compute, OpenStack and storage to help scale SPICE.
The Met Office has selected three HPC vendors to provide high performance computing capabilities for its new Scientific Processing and Intensive Compute Environment (SPICE) system.
Known as a world leader in weather and climate forecasting, SPICE will be used to enable weather and climate researchers to dramatically reduce time required to analyse massive amounts of climate simulation data. The need for a new HPC system was driven by the Met Office need for a powerful system for post-processing data and analysis downstream of the primary HPC facility.
In order to help its 500 scientists conduct research using data-intensive, high-resolution models to increase forecast accuracy and provide a deeper understanding of climate change, the Met Office brought on board SGI, Bright Computing and DataDirect Networks (DDN).
SGI will be powering the SPICE initiative, as well as upgrading the Met’s Managed Archive Storage System (MASS).
To support the growth in its MASS which is a critical adjunct to the Met Office’s supercomputer system archive, the Met Office selected SGI’s solution with DDN storage. MASS acts as a repository or archive for the data resulting from scientific research carried out on the supercomputer as well as global observational data. By 2020, this crucial storage archive is predicted to grow to about 300 Petabytes of weather and climate research data.
The Met Office will be integrating Bright Computing’s Cluster Manager for HPC to deploy the new SPICE cluster over bare metal, providing single-pane-of-glass management for the hardware, operating system, HPC software, and users. The Met Office also chose to install Bright OpenStack to enable the IT team easy deployment, provision, and management of its OpenStack-based private-cloud infrastructure.
The three vendors have come together to give the Met Office a combined solution of compute, OpenStack, and storage, which will enable the scaling of SPICE predictably while delivering high-throughput performance to handle simultaneous data reads/writes.