Ship-based cloud computing for advancing oceanographic research capabilities

The use of advanced high performance computing (HPC) or so-called cloud computing systems aboard ships is a new capability. We discuss some of the ways oceanographers and marine scientists across a variety of disciplines are working to understand how they can best utilize shipboard HPC capabilities for data-intensive operational oceanography. Fundamentally, a comprehensive understanding of the ocean's complex biophysical and geochemical dynamics relies heavily on numeric modeling. While oceanographic data are collected at sea, their assimilation into models conventionally takes place ashore, where computing resources are readily available. Due to limitations of data transmission cost and bandwidth, this approach creates delays with data integration and makes it hard to adjust the course of spatial and GIS-based operational oceanographic observations in response to insights from the incoming data. To shorten feedback between observations and modeling, and provide researchers with better visibility into observed processes through visualization and dissemination of data via Cloud Services, Schmidt Ocean Institute (SOI) installed a high performance computing (HPC) system on RV FALKOR (Fig. 1) in January of 2015, making her the first research vessel with a supercomputing system freely available to the global oceanographic community.