High Performant File System Workloads for AI and HPC on AWS using IBM Spectrum Scale
Author | : Sanjay Sudam |
Publisher | : IBM Redbooks |
Total Pages | : 34 |
Release | : 2021-03-31 |
ISBN-10 | : 9780738459554 |
ISBN-13 | : 0738459550 |
Rating | : 4/5 (54 Downloads) |
Book excerpt: This IBM® Redpaper® publication is intended to facilitate the deployment and configuration of the IBM Spectrum® Scale based high-performance storage solutions for the scalable data and AI solutions on Amazon Web Services (AWS). Configuration, testing results, and tuning guidelines for running the IBM Spectrum Scale based high-performance storage solutions for the data and AI workloads on AWS are the focus areas of the paper. The LAB Validation was conducted with the Red Hat Linux nodes to IBM Spectrum Scale by using the various Amazon Elastic Compute Cloud (EC2) instances. Simultaneous workloads are simulated across multiple Amazon EC2 nodes running with Red Hat Linux to determine scalability against the IBM Spectrum Scale clustered file system. Solution architecture, configuration details, and performance tuning demonstrate how to maximize data and AI application performance with IBM Spectrum Scale on AWS.