Telecommunications

Transforming the Telecom Industry with Accelerated Networking for 5G and Edge AI

Enhance Agility with Cloud-Native NFV

Advances in Cloud technologies, new service offerings, and the ever-increasing demand for bandwidth are transforming communication service providers (CSPs). To keep pace, packet processing must be offloaded to hardware using DPDK, and the NVIDIA ASAP2 (accelerated switching and packet processing) technology.

ASAP2 provides a time-bound packet flow engine that enables software-defined, hardware-accelerated infrastructure to provide optimized performance for virtual network functions (VNF) and containerized network functions (CNF). With ASAP2, I/O processing is offloaded to the NIC, freeing up server CPU. The server can dedicate its resources to application processing, resulting in higher system efficiency.

Intelligent Edge AI Computing

Automated vehicles, AR/VR, cloud gaming, Industry 4.0. These and many other emerging use cases have in common a need for Edge Artificial Intelligence (AI): the nexus of IoT, 5G, and AI running on multi-access edge computing (MEC) equipment.

Edge AI on MEC allows processing to be moved away from centralized data centers, so processing occurs closer to the user for optimal latency. NVIDIA offers a complete portfolio of ConnectX^® SmartNICs and BlueField^™ DPUs to offload networking and security functions and accelerate MEC packet processing using ASAP2 to implement AI decision making closer to the edge.

Bringing 5G Services to Life

With 5G, CSPs are adopting cloud technologies to build open, cost-effective, scalable and programmable networks. The NVIDIA 5T for 5G technology provides an ideal reference architecture for software-defined, hardware-accelerated 5G CloudRAN (Cloud radio access networks). 5T for 5G offers latency-sensitive scheduling with precise time accuracy of just 16ns to help solve the complex scheduling problems of 5G. This approach adds flexibility and efficiencies allowing CSPs to transform their business model and provide scalable data services to tenants and sub-tenants that offer elasticity and demand-based cost structures.

Read Blog

Virtual RAN and Edge Platforms

Turbo charge your network edge to meet the challenges of 5G data processing for next-generation software-defined and AI infrastructure with NVIDIA Virtual RAN and Edge Platforms.

Learn More

Cloud & Telco Software Ecosystem

Essential for high-performance computing, artificial intelligence, and big data, NVIDIA Networking InfiniBand is the clear leader that introduces capabilities into the network that are specifically designed to achieve maximum performance. With In-network computing capabilities, such as SHARP™ actually transforms the network into a powerful compute resource which can process data as it traverses the network. A wide range of native hardware-based offloads including MPI tag-matching and rendezvous protocol offloads, GPUDirect RDMA and NVMEoF are just a few examples of advanced technologies that are readily employed by NVIDIA InfiniBand. Network reliability with NVIDIA Networking SHIELD™ improves resiliency by speeding up fault recovery by 5000 times and enables autonomous self-healing capabilities.