Amax Engineering Launches StorMax 200Gb/s AFA and Hybrid Flash Solutions With NVMe-oF on IB

Amax Engineering‘s HPC and AI solutions group announced an updated high-performance storage solution series, StorMax, all-flash and hybrid flash storage solutions featuring 200Gb/s NVMe-oFon IB with Mellanox ConnectX-6 adapters.

Stormax A 1110nv

StorMax platforms are performance, secure and scalable architectures, with price-performance storage capability that accelerates all AI computing, database, big data analytics, cloud, web 2.0, and video processing data.

As an Elite Nvidia partner, the company designed the StorMax performance storage series to pair with the Nvidia A100 GPU compute solutions combined with Mellanox 200Gb/s networking to supply compute and storage I/O through streamlined parallelism, performance and scale.

“With the availability of Mellanox 200Gb/s fabrics for GPU compute and storage as the new standard connectivity for server platforms hosting Nvidia A100 accelerator cards, we see an immediate need for high bandwidth low latency NVMe storage solutions matching GPU server bandwidth,” said Dr. Rene Meyer, VP, technology and product development “Maximizing storage bandwidth will reduce GPU idle time, while shared storage allows for more effective management of training data lakes for parallel training in GPU cluster and multi-user settings. We are excited to announce our 200Gb/s NVMe stand-alone and scale-out storage solutions as part of our Nvidia DGX POD turnkey offerings.“

Performance storage software logical architecture

The StorMax series features ConnectX-6 with ‘Virtual Protocol Interconnect’ offering 2 ports of 200Gb/s IB and Ethernet connectivity, sub-600 nanosecond latency, and 215 million messages per second. ConnectX-6 cards bring innovation to network security by providing block-level encryption; off-loaded by the ConnectX-6 hardware, improving latency and saving CPU cycles.

In addition, the company’s StorMax 200Gb/s storage systems deliver low-latency distributed block storage for web-scale applications, enabling shared NVMe across any network and supports any local or distributed file system. It features a management layer that abstracts underlying hardware with CPU offload, creates logical volumes with redundancy, and provides centralized, management and monitoring. All applications benefit from the low latency, high throughput and high IO/s of a local NVMe device with the convenience of centralized storage while avoiding proprietary hardware lock-in and reducing the overall TCO.

Stormax A 2440

StorMax series 200Gb/s systems include:

StorMax A-1110NV: 1U single-socket AMD EPYC processor all-flash storage server, with 10×2.5″ hot-swap NVMe U.2 drive bays
StorMax A-2440L: 2U 4-node single-socket AMD EPYC processor storage server, with 24×2.5″ hot-swap NVMe U.2 drive bays (6 bays per node)

StorMax storage solutions coupled with Nvidia A100 GPU servers are for companies tackling today’s compute challenges. Customers have simple access to supercomputing infrastructure that delivers performance and scalability to accelerate the largest AI workloads.

Read also:
Amax: Storage Servers With AMD EPYC 7202 Series Processor
Including ServMax with up to 128 cores (256 threads), 8TB memory, and 5+ TFlops of double-precision performance, StorMax up to 360TB NVMe flash storage with 80GB/s IO bandwidth per server, and BrainMax with up to 6x GPU direct host-device connection
August 15, 2019 | Press Release
Amax Deep Learning and Storage Appliance Platforms With Xeon Scalable Processors
Including StorMax I/O optimized storage servers, hyper-converged cluster-in-a-box servers, and CloudMax cloud server
April 11, 2019 | Press Release