Case Study: Federal Government Agency achieves high-performance facial recognition analytics with Pavilion Data Systems

Federal Government Agency - Customer Case Study

The Federal Government Agency needed a storage solution for a large-scale facial recognition and AI/ML analytics environment, with requirements for very high throughput, over 10M IOPS, low latency, strong encryption, and limited rack space. Pavilion Data Systems evaluated the agency’s plan for a massive IBM Spectrum Scale (GPFS) cluster and identified its HyperParallel Data Platform as the best fit for the workload.

Pavilion Data Systems implemented two Pavilion HyperParallel Data Platforms with GPFS, using NVMe-oF to deliver local-NVMe-like parallel performance along with enterprise storage features. The solution delivered over 150 GB/s of bandwidth, supported linear scalability for growth, and achieved 10x lower TCO and 2x higher compute density versus alternatives, while also improving manageability and meeting the agency’s performance goals.