NVIDIA EGX Scales Real-Time Edge AI from Jetson Nano to T4 Server Racks

May 28, 2019 Brandon Lewis

COMPUTEX. NVIDIA announced the EGX accelerated computing platform at Computex Taipei 2019, an edge-centric artificial intelligence (AI) solution based on the company’s Jetson Nano, T4 servers, Edge Stack, and Red Hat’s OpenShift Kubernetes container orchestration platform. The scalable offering also integrates networking, storage, and security technologies from Cisco and Mellanox to enable real-time AI processing on continuous data streams

Designed as a low-latency, high-throughput, and deterministic edge AI solution that minimizes the need to send data to the cloud, NVIDIA EGX is compatible with hardware platforms ranging from the Jetson Nano (5-10 W power consumption, 472 GFLOPS performance) to a full rack of T4 servers capable of 10,000 TOPS.

On top of these hardware solutions, NVIDIA has layered on its Edge Stack, which contains drivers, a CUDA Kubernetes plugin, CUDA container runtime, CUDA-X libraries, and containerized frameworks such as TensorRT, TensorRT Inference Server, and DeepStream. A partnership with Red Hat has resulted in an optimized integration of Edge Stack and OpenShift, which enables smooth container management for distributed operations.

Mellanox Smart NICs and Cisco Ethernet/IP networking, compute, storage, fabric, and software offerings help facilitate consistency across edge to cloud workloads. Remote management of NVIDIA EGX is available from AWS IoT Greengrass and Microsoft Azure IoT Edge.

EGX servers are available from embedded and IoT OEMs including:

  • Abaco
  • Acer
  • Advantech
  • ASRock Rack
  • ASUS
  • AverMedia
  • Cisco
  • Cloudian
  • Connect Tech
  • Curtiss-Wright
  • Dell EMC
  • Fujitsu
  • HPE
  • Leetop
  • MiiVii
  • Musashi Seimitsu
  • QCT
  • Sugon
  • Supermicro
  • Tyan
  • WiBase
  • Wiwynn

More information is available at www.nvidia.com/en-us/data-center/products/egx-edge-computing/.

About the Author

Brandon Lewis

Brandon Lewis, Editor-in-Chief of Embedded Computing Design, is responsible for guiding the property's content strategy, editorial direction, and engineering community engagement, which includes IoT Design, Automotive Embedded Systems, the Power Page, Industrial AI & Machine Learning, and other publications. As an experienced technical journalist, editor, and reporter with an aptitude for identifying key technologies, products, and market trends in the embedded technology sector, he enjoys covering topics that range from development kits and tools to cyber security and technology business models. Brandon received a BA in English Literature from Arizona State University, where he graduated cum laude. He can be reached by email at brandon.lewis@opensysmedia.com.

Follow on Twitter Follow on Linkedin Visit Website More Content by Brandon Lewis
Previous Article
OmniVision’s New OV0VA10 SoC Now Available

SoC uses advanced pixel technology, featuring compact size, top VGA camera performance and maximum notebook...

Next Article
The Quantum Menace
The Quantum Menace

Effective quantum computing – when it actually gets implemented – will have profound and far-reaching impli...