Affirmed & Intel Enable Cost-Effective and High Scale 5G Core
Affirmed Networks’ 5GCN cloud-native network function reached 200 Gbps on 100 GbE links using Intel® Xeon® Scalable processors and Intel® FPGA-based Programmable Acceleration Card and Dell EMC’s PowerEdge server
5G technology promises to greatly enhance the speed, coverage, and responsiveness of wireless networks and will drive new use cases such as interactive television, high-definition and 3D video, social gaming, virtual and augmented reality (VR and AR), robotics, automated vehicles, advanced manufacturing, and healthcare imaging. Some of these use cases require high bandwidth, while others require low latency. A new type of platform is needed to deliver the best of IT while meeting mobility network requirements. This new infrastructure will need to be cost-effective, flexible, easy to deploy and scale efficiently with growing traffic and services. These conditions can be met through an optimized, joint hardware and software solution using the leading COTS x86-based platforms and hardware acceleration by FPGA.
Intel, Affirmed Networks, and Dell EMC have developed a proof of concept for high-performance Next-Generation Core Networks (NGCN). This white paper presents a state-of-the-art solution for the commercial, standalone 5G core network (5GCN) that demonstrates the 100 Gbps/CPU socket (200 Gbps in the next-gen solution) using a 100 GbE NIC interface and Intel® Xeon® Scalable processor-based server. The combination of software optimization and hardware acceleration resulted in the phenomenal 20 Gbps/Core.
Smart load balancing using Intel® FPGA and CPU cache optimizations with Affirmed Networks’ revolutionary 5G User Plane Function (UPF) enhances software performance of 5GCN applications. Intel FPGA acceleration addresses traditional I/O limitations, allowing termination of 100 GbE interfaces per CPU socket, 20 Gbps per core, and enabling 16 or more worker cores for packet processing. Get the White Paper to read more…
Download the Intel White Paper