HPC Infrastructure to Support the Next-Generation ARM Facility Data Operations

Description

As part of its Next-Generation ARM Decadal Vision, ARM is implementing two major objectives: 1) establish observation-modeling megasites at the Southern Great Plains (SGP) and North Slope of Alaska (NSA); and 2) produce routine high-resolution large-eddy model simulations at ARM sites to improve observational retrieval and parameterizations of clouds, aerosols, and radiation in climate models. New cloud and precipitation radar instruments have been added at ARM sites to support these activities. With the induction of additional radars at all ARM sites, data operations anticipate the incoming data volume archived to grow from 30 TB per month to 160 TB per month. In addition, the Large-Eddy Simulation (LES) ARM Symbiotic Simulation and Observation (LASSO) workflow—lead by Dr. William Gustafson (Pacific Northwest National Laboratory) and Dr. Andrew Vogelmann (Brookhaven National Laboratory)—will conduct routine LES modeling at the SGP megasite. Currently, the LASSO project is in the pilot phase and routine operations are expected to begin in mid-2017. This project will produce a library of simulations and tens of TB of data outputs daily. These initiatives require a next-generation computing infrastructure to support model simulations, peta-scale storage, and big-data analytics for successful ARM climate science research. The ARM Data Center has designed a high-performance facility architecture that integrates multiple levels of computing resources. In this poster, authors describe upcoming high-performance computing infrastructures in support of the Next-Generation ARM Facility.