Inventec‘s Edge server, E850G4 is certified for functionality and performance of the GPU Optimized AI stack from NVIDIA, security, and remote management. The 2U GPU server, targets multi-access edge computing (MEC), content acceleration, network-in-a-box, and network function virtualization (NFV) applications, among others, in the booming new market of 5G and edge datacenter.
NVIDIA EGX Edge AI platform
With the future pointing towards convergence of IoT and AI at the edge, allowing all industries to deliver automated intelligence from billions of sensors at the point of action, a scalable, accelerated platform that can handle continuous delivery of information securely and drive decisions in real-time is required. NVIDIA unveiled the latest NVIDIA EGX platform capabilities and NGC features for deploying accelerated AI at the edge and in the data center at GTC Digital in May 2020.
NVIDIA EGX platform enables companies to perform low-latency and high performance AI inference at the edge and includes GPU-optimized AI software from NVIDIA NGC, a registry of containers, pre-trained models, and helm charts along with a private registry that simplifies building, collaborating, and deploying AI software securely at the edge.
OTII-based Edge computing server
The E850G4 server system is encased in a computing box with a NUMA (non-uniform memory access) balanced design for multiple SKUs for acceleration and storage scenarios. This system is also a China Open Telecom IT Infrastructure (OTII) server design for WW5G edge computing applications such as multi-access edge computing, and fog computing. In addition, E850G4 offers enhanced serviceability and reliability including:
- Up to 4x NVIDIA T4 Tensor Core GPUs
- Dual Socket: 2nd Gen Intel Xeon scalable processors (Cascade Lake)
- 12x DIMM: Up to 1.5TB 3DC ECC DDR4-2933MHz, RDIMM/LRDIMM and Intel Optane DC Persistent Memory (DCPMM)
- Onboard 2x 10G Ethernet port
- Up to 7x PCIe Gen3 slot to support multi-scenarios
- 6+(2) SFF SAS/SATA/ NVMe hot-plug drive
- Tool-less design of serviceable parts
- Optimized TCO