Arm’s new Neoverse platforms boost compute and throughput for 5G

Arm is preparing for a post-Moore's-Law world where compute needs to be more heterogeneous. (Getty Images)

Arm is known for its chipset intellectual property (IP), which is widely used in mobile devices as well as network infrastructure. Recently, the company has been making a concerted effort to gain even more market share in the network infrastructure space because it predicts the ecosystem is going to undergo significant change as the number of end devices proliferate.

The company predicts that the number of IoT and other endpoint devices will, ultimately, far surpass the current number of global smartphones. To meet the demand, the company unveiled in late 2018 its Arm Neoverse brand for high-performance computing and storage in network infrastructure.

And today Arm announced its Neoverse N1 and E1 platforms. Neoverse N1 is a high-compute oriented product optimized for 7nm technology. And Neoverse E1 is designed for driving very fast data throughput.

Sponsored by Ciena

Because you asked. Adaptive IP™

There’s a new way to modernize and expand your IP-based networks—from access to metro—that’s automated, open, and lean.

Potential use cases for N1 would be to enable data to be quickly onboarded and processed at a 5G base station or backhauled to a data center where the computing could then be done.

The E1 platform is designed to handle increased throughput demands that are emerging, said Drew Henry, SVP and general manager of Arm’s infrastructure line of business, during a media briefing. He said enterprises need increased throughput to move high bandwidth data around the network. Arm designed the Neoverse E1 platform especially to enable the transition from 4G to a more scalable 5G infrastructure with more diverse compute requirements. E1 also delivers scalable throughput for edge to core data transport.

“We’re focused on solutions for the global internet infrastructure from hyperscale data centers all the way out to the onboarding spots with edge servers and 5G base stations,” said Henry.

He added that networking infrastructure is moving from general purpose computing and processors to heterogeneous computing and that in a post-Moore’s-Law world, customers need flexible designs to allow them to build compute, storage and networking with processors that meet their specific needs. “It’s why we’re seeing such remarkable things in the silicon ecosystem with very large companies going off and building their own silicon,” he said.

RELATED: Changes underway at ARM

For its Neoverse platform, Arm not only provides the CPU IP, but it also offers the architecture and reference designs. “We provide recommendations about how to design, and we look at other pieces of IP that might be needed and validate IP from other providers so they work well with our infrastructure,” said Henry. Arm is also working with standards and certification groups as well as open source groups.

Henry said that the infrastructure business is Arm’s “fastest growing business unit,” and the company expects it to be one of its largest contributors to growth over the next five years.

Suggested Articles

Amazon is putting the loss of the $10 billion Pentagon JEDI cloud contract squarely on the shoulders of President Trump, according to a filing.

Ericsson announced late Friday afternoon that it had agreed to pay $1.06 billion in fines to U.S. regulators after pleading guilty to bribery charges.

Amid a reorganization effort, Cisco is losing two key executives, including a former CIO and its top data center sales leader.