With the move to 5G, so too comes the next generation of hardware and software for network infrastructure. In fact, the next wave of network transformation represents a $25 billion silicon opportunity by 2023, as industries look to benefit from 5G, edge, and AI (artificial intelligence).
“We see tremendous momentum and opportunity in … transforming the radio-access network and bringing to life 5G, edge, and artificial intelligence, and just pushing toward that network transformation,” says Cristina Rodriguez, vice president, Data Platforms Group, and general manager, Wireless Access Network Division, Intel.
Intel recently announced an expanded lineup of hardware, software, and solutions for network infrastructure, which comes after a longer journey with the communications industry to transform the network with a software-defined, agile, and scalable infrastructure.
“As you probably know, we have a long history of working on virtualization. This started at the core more than a decade ago,” says Rodriguez. “We see the same benefit be applied to RAN (radio-access network)—flexibility, scalability, agility, cost-effective solutions.”
This next wave of network transformation is being driven by cloud-native technologies and virtualization, with an expected 50% of core network deployments transforming to virtualized networks this year, and virtualization extending to the RAN.
“At a very high level, there are two ways to deploy radio-access networks,” Rodriguez adds. “One is the traditional way that is based on custom solutions and hardware that is optimized for the specific solutions. And then there is what is now taking a lot of momentum and having a lot of attention in the industry, which is deploying the radio-access network on a server architecture—or virtualized—where we are going to take functions that used to be running hardware; we are taking the solutions and running in software.”
Intel’s expanded lineup of hardware, software, and solutions for network infrastructure includes: enhancements to Intel’s software reference architecture, FlexRAN; Intel’s virtualized vRAN (virtualized radio-access network) dedicated accelerator; network-optimized next-generation Intel Xeon Scalable and D processors (code-named Ice Lake); and upgraded Intel Select Solutions for NFVI (Network Function Virtualization Infrastructure).
FlexRAN grew to nearly 100 licenses and added enhancements including optimizations to its massive multiple input, multiple output mid-band pipeline for increased bandwidth and support for ultra-reliable low-latency communication.
The low-power and low-cost acceleration solution for vRAN deployments is based on Intel eASIC technology and it offloads and accelerates the computing-intensive process of forward error correction. This frees up more processing power within Intel Xeon processors for channel capacity and edge-based services and applications. The ACC100 is sampling to customers now.
With the next-generation Intel Xeon processors for network infrastructure, customers can use a common architecture across the network for various workloads and performance requirements.
To improve application efficiency and network performance for high-performance network workloads, Intel Select Solutions for NFVI Red Hat, NFVI Ubuntu, and NFVI Forwarding Platform have been upgraded to support the new Intel Ethernet 800 Series Network Adapter (code-named Columbiaville), which delivers increased performance and DDP (dynamic device personalization) to maximize platform performance. These upgraded solutions are coming soon.
With virtualization, customers are able to realize economies of scale and are able to be more agile. “We have more than 15,000 end customers edge to deployment,” says Rodriguez. “We help customers deliver better business and bring value to them.”
Going forward, these transformations in the network will help achieve the full potential of 5G, edge, and AI across industries such as retail, education, healthcare, and more. “RAN innovation is taking off and it is a very exciting time in both the traditional and the virtualized,” says Rodriguez.