Intel today revealed a mixture of its hardware and software for radio access network (RAN) and edge computing called the Intel Network Platform.

The architecture includes processors, accelerators, Ethernet adaptors, memory, software toolkits and blueprints traversing the core, access, and edge of networks.

Intel is mostly known as a CPU vendor, but the company is increasingly investing at the platform level by integrating processors, other components, and software to meet the goals of communications service providers, Dan Rodriguez, VP and GM of Intel’s Network Platforms Group, explained in an interview with SDxCentral. 

“No other silicon provider offers this type of flexible, comprehensive software environment that really jumpstarts our customers’ developments, and really helps them ensure that they can deliver products that meet their customers’ needs with the right performance, the right latency, and then of course the right agility and flexibility,” Rodriguez claimed.

Intel is also targeting emerging opportunities in enterprise, an interest shared among network operators that are upgrading network technology and pressed to develop new lines of business riding on those networks. 

“We’re really seeing the convergence of operational technology, IT, and communications technology all come together to really benefit the enterprise,” Rodriguez said. “We’re trying to pull together all of our silicon ingredients with a robust software platform that really is a horizontal platform, but also has vertical-specific recipes” for industries such as industrial IoT, health care, and utilities. 

Intel is pooling artificial intelligence (AI) algorithms, its FlexRAN reference software for virtualized RAN (vRAN), and two variations of Intel Smart Edge to present operators with a comprehensive and customizable architecture for 5G, edge computing, and enterprise-specific requirements.

Intel Smart Edge, which targets the multi-access edge computing (MEC) layer, includes commercial software for enterprise use on premises, such as private networks, and Intel Smart Edge Open (formerly known as OpenNESS). “We are unifying the Intel Smart Edge and openNESS for greater customer flexibility, whether they want to buy a commercial solution or build their own solution,” Rodriguez said.

Software for mobile edge computing allows operators to “abstract the network and really scale it across a multi-access network that could include WiFi, wired, 4G, or 5G. So it needs to onboard and authenticate all sorts of virtual network functions,” he said. “It also has a tough job of onboarding and authenticating all sorts of new AI-based compute applications scaling across those multiple market segments.”

Red Hat is using both variants of Intel Smart Edge in OpenShift, the former’s Kubernetes-based platform that operators are using for 5G RAN and network cores. Dell is also partnering with Intel on the Smart Edge commercial offering in a bid to bolster its position in the telecom market. 

Intel is also confident that its third-generation Xeon Scalable processors, especially those with telco-specific traits, will better target 5G network transformation efforts in a more flexible manner, according to Rodriguez. 

The company’s far-reaching activities in the telecom market are underpinned by Intel’s silicon, accelerators, and FlexRAN software that pulls multiple resources together and provides operators with a blueprint for vRAN deployments.  

“When you think about that ability to innovate in silicon, innovate in software, and then make it easy for our customers to use Intel solutions, that’s what really differentiates Intel in the market,” Rodriguez said.

Half of all core network deployments in 2020 were virtualized and Intel is the leading network silicon provider there. The company also enjoys a near monopoly on vRAN deployments due to its early push on FlexRAN with a large group of partners. 

“Nearly all commercial deployments are running on Intel technology today,” Rodriguez said, adding that there are more than 100 licensees of FlexRAN today. “When you think about that and the investments we’re continuing to make, that lead matters because the more proven a solution is, the more mature the ecosystem is, the easier it is for our customers to use that solution,” he said.

Rodriguez acknowledges that competition in that area is growing, but he said Intel is confident that it has the right strategy, technology, and ecosystem of partners to maintain its market lead. 

“You have to really innovate in a way that you can control that technology innovation, and then scale that innovation,” he said, claiming that creating a more diverse pool of vendors, one of the primary goals of vRAN and open RAN, is achieved via FlexRAN.

Multiple server vendors, cloud automation software providers, and chipmakers are at various stages of development in creating platforms for vRAN. “In the end, the service providers are able to select the right recipe that meets their needs,” Rodriguez said. "Yes, competition’s out there. That’s just more of a reflection that the total addressable market is real and vRAN is going to happen.”

While FlexRAN has some attributes that allow for open use, enabling operators to mix and match certain components or technologies, silicon remains a critical piece of gear that is restricted exclusively to Intel in a FlexRAN architecture. If an operator chose to go use silicon from Qualcomm, which plans to provide its own vRAN and open RAN technical blueprint next year, FlexRAN would not be available. The same goes for RAN-centric silicon under development from other chipmakers, at least for now.

“FlexRAN allows you to create this robust ecosystem,” Rodriguez countered. Beyond that, however, “there’s nothing I can really share at this time, and there’s no plans to share.”