WekaIO: 5 Reasons Why NAS Can’t Support Deep Learning Workloads
Recorded Aug 14 09:30 pm
About this Webinar
Legacy storage systems, like NAS, were architected when spinning disk and slower networking technologies were the industry standard. In this webinar, we’ll present five reasons why NAS can’t keep pace with the I/O demands of new deep learning workloads. To support these workloads, the data processing layer has to have immediate access to, and a constant supply of, data. Here NAS falls short, because the data gets bottle-necked between the compute and storage. WekaIO Matrix™ is a next-generation shared, distributed file system that visualizes the SSDs into one logical pool of fast storage presenting a global namespace to the host applications. Matrix was written from scratch to leverage the benefits of standard Intel x86 architecture combined with NVMe. The result is an easy to deploy, easy to manage storage architecture that is a radical departure from legacy NAS systems. By optimizing Matrix for flash, the storage solution is ideal for deep learning and high-performance computing workloads.
David Hiatt, Director of Product Marketing, WekaIO