High-Performance Computing (HPC) and scalable Artificial Intelligence (AI) workloads typically use parallel file systems to make sure that all the computing nodes in a cluster can directly access the data they need without being bottlenecked by control nodes. Organizations running these workloads are looking to more easily include external data sources as well as, of course, improve performance. Making the correct infrastructure choice is often a major contributor to the success or failure of projects leveraging large scale data.
In our latest LightBoard video, George Crump, Lead Analyst at Storage Switzerland and Sven Oehme, Chief Research Officer at DDN, discuss how DDNs EXAScaler clustered storage solution, powered by its new EXA5 Filesystem, meets these demands. Watch the video to learn more about EXA and how, in its latest release, EXA5 provides support for NFS, SMB, and S3 clients to support edge and IoT devices. In the video, we also discuss GPUDirect to Storage, a collaboration between NVIDIA and DDN which enables a direct path between storage and GPU memory.