P40 Enhancing
community collaborations through NWP software containers
Harrold, Michelle, Jamie Wolff, John Halley Gotway, Kathryn
Fossell, National
Center for Atmospheric Research (NCAR) and Developmental Testbed
Center (DTC), and John Exby, NCAR
A frequent stumbling block when first running a modeling
system is properly setting up and compiling all of the necessary code
components, including a number of external libraries. In addition to running
a forecast model, users often need pre- and post-processing software as well
as a means to visualize and verify output from their model runs. To ease the
burden of setting up a new system from the ground up, the concept of
"containers" has quickly been gaining traction in the modeling community.
Containers allow for end-to-end software systems to be bundled and provided
to users, including the operating system, libraries, and code. This
eliminates a myriad of technical issues frequently encountered when first
spinning up on compiling all of the necessary components of numerical weather
prediction systems (NWP). Containers also provide an avenue
to run experiments using the same platform and code bases, naturally lending
itself to easy sharing and comparing of results among collaborating groups,
and enables the ability to reproduce results. |