CC* Storage: Implementation of a Distributed, Shareable, and Parallel Storage Resource at San Diego State University to Facilitate High-Performance Computing for Climate Science

点击翻译按钮获取中文摘要

This project develops a high-capacity, high-performing storage system closely matched to the identified needs of climate science, but also of interest to a wide range of research and engineering fields such as brain and behavior research, high-energy physics, renewable energy, genomics, bioinformatics, computational chemistry, and the study of dark matter. The two-tiered design of the storage system is consistent with best practices and would maximize research capacity and utilization.The effort would create a distributed, shareable and parallel cluster running the BeeGFS file system (a European development), delivering about 1.8 PB (usable) storage and capable of up to 5GB/sec of streaming performance. The project would support twelve science, technology, engineering and mathematics (STEM) projects in a broad range of disciplines; co-investigators in mechanical, environmental, civil and aerospace engineering, computer science, and biology are participating. The results would improve input-output operations associated with a number of research and engineering computation bottlenecks: parallel sequential writing during distributed-parallel geochemical, computational fluid dynamics, and structural mechanics simulations; new file creation during real-time data acquisition; and random read operations during post-processing and visualization.