Quanta Cloud Technology (QCT), a global data center solutions provider, has formed a strategic partnership with NVIDIA. The two companies collaborate on a new integrated architecture based on NVIDIA’s EGX Edge Computing Platform optimized for AI applications, as employed in 5G and other of today’s most advanced industries. The cooperation is aimed at helping data center customers to immediately identify systems capable of running Artificial Intelligence (AI) workloads, thereby eliminating the typically time-consuming, cumbersome validation process for AI-ready systems.
With the rise of 5G technology, where data is processed "in motion" directly when it is generated or received, local computing capacity is becoming increasingly important. As a result, AI supported applications - whether related to IoT, streaming of sensor data, or cloud-native features - will also move closer the edge. The same applies for AI capable servers from QCT that are now shifted towards the end-user, where large data volumes are received and produced on a daily basis.
NVIDIA’s NGC-Ready Program Enhanced for Edge
NVIDIA’s scalable, GPU accelerated EGX Edge Computing Platform provides superior and stable performance, regardless if the workloads are located on-premise, in the cloud, or at the edge. The system supports NVIDIA’s NGC-Ready-for-Edge program that quickly identifies systems capable of running advanced AI workloads at the edge.
QCT’s NGC-Ready Computing Platforms
QCT’s and NVIDIA’s AI-ready infrastructure optimizations required to facilitate the handling of AI workloads represent an important step not only for 5G, but also for various vertical markets, including retail, manufacturing, healthcare, or smart cities. Using NVIDIA’s EGX Computing Platform enables QCT to deliver dedicated server solutions for the edge that allows for smooth and flexible handling of AI applications.
With the QuantaGrid D52BV-2U, QuantaGrid SD2H-1U, QuantaGrid D43K-1U and QuantaGrid S43KL-1U, QCT currently offers four platforms that have been validated with the latest NGC-Ready guideline. Additional models will be released through the end of 2019. Each of these systems is able to support NVIDIA T4 or V100 Tensor Core GPUs. NVIDIA features, including driver, container runtime, Kubernetes NVIDIA device plug-in and monitoring plug-in, can now be managed centrally with the cloud-native NVIDIA GPU Operator, making AI applications much easier to deploy. Via a Helm chart that can be quickly installed on a QCT system, it is now possible to run the setup of a GPU-powered node fully automatically.
