Exclusive: HPE Networking advances 'self-driving' networks

HPE Networking says the era of the 'self-driving' enterprise network is approaching, as organisations grapple with AI workloads, hybrid work and mounting security threats.

Andrew Fox, who formally took on the role of Country Manager for Australia and New Zealand after 18 months in an acting capacity, is now leading the integration of HPE's Networking's assets following the Juniper acquisition.

From January, the combined operation has been trading simply as HPE Networking, bringing together teams previously operating under Aruba and Juniper brands.

"It's fantastic to be in the job and fantastic to have the job of combining two talented teams, although that comes with an awful lot of moving parts," said Fox, Country Manager, Australia & NZ, HPE Networking. "They are two amazing teams and culturally similar. I'm having a ball trying to bring them together into one cohesive team. From the first of January, we are just one, just HPE Networking. There's no Aruba, there's no Juniper."

AI operations

Fox argues that enterprise networking has shifted from being "plumbing" to a strategic platform underpinning security, performance and user experience. The most significant change now, he says, is the application of AI to network operations.

"The single biggest disruptor, and the single biggest opportunity, is what we can do inside of networking with AI specifically. I'm not talking about building networks for AI clusters, which we absolutely do, but using AI to better run a network," said Fox. "This concept of AIOps, or AI operations, is probably the biggest trend that's front and centre with our customers at the moment."

With organisations facing skills shortages and increasing complexity, Fox said AI-driven automation can identify and remediate issues before users raise tickets.

"If you could see that, recognise it, determine what you need to do to fix it, and then implement that in real time, that's gold," he said. "The platform can do that quickly, and you don't have to have a team scratching their head and trying to figure out what's going on."

Alongside AI for networking, Fox highlighted networking for AI as a parallel growth area. Large AI clusters require vast bandwidth and low latency to keep GPUs supplied with data.

"We're now selling 800 gigabit per second connections, and 1.6 terabit per second is just about here, and then it'll be 3.2 terabits per second, which are terrifyingly fast," he said. "When you think about where we've come from, one gigabit Ethernet we're used to plugging into the side of our PCs, this is 800 times that fast."

Unified security

Security is becoming inseparable from networking architecture, according to Fox, with customers moving away from perimeter-only models towards integrated controls at every layer.

"The best way to keep an organisation safe, or a user safe, is to have security at every point," he said. "If you can stop them at the access point, stop them at the switch, stop them everywhere, then in theory that's just easier to keep organisations safe."

He described this approach as integral security, embedding policy enforcement across infrastructure rather than relying solely on edge firewalls.

At the same time, enterprises are consolidating network domains that were previously managed in silos.

"We're seeing a trend of all of those coming together," Fox said. "If you can have one platform, one set of skills, one team, and manage all the pieces of the network that matter, the payback on that is crazy good."

AI operations strengthens that model, he added, reducing the need for large operational teams. He pointed to large retail deployments where centralised management allows thousands of access points to be rolled out with minimal on-site expertise.

"You don't need a legion of people, because you've got a platform that will do the heavy lifting for you, powered by AI," he said.

Hybrid pressure

Hybrid working continues to reshape network and security requirements, particularly across Australia and New Zealand where distributed workforces are common.

"When all your people are in one place, it's really easy to connect and secure them," Fox said. "But as soon as they start picking up and one's at home, one's in a coffee shop, one's at a customer, then it's much harder to secure and to keep them safe."

Technologies such as secure service edge (SSE), applied at the device level, are becoming more important in maintaining consistent policy regardless of location.

Edge computing is also rising in importance as organisations seek to process data closer to where it is generated.

"You need fresh data," Fox said. "We're collecting millions of data points on the user experience every user is having. It's so much data that you can't afford to send it all back to a central place."

Processing data at or near the edge reduces latency and cost, while enabling faster business decisions.

"If you can make a quick decision that might help your business go faster out near the edge, that's a beautiful thing," he said.

Skills shift

As automation increases, Fox expects traditional command-line configuration skills to decline in importance.

"The skills required to sign into each individual network appliance and configure that on the command line, that will be gone," he said. "In the main, already is."

Instead, he believes architecture, policy design, governance and security expertise will define future network roles.

"It's having the clarity around the policy and the process and how things work. That's going to be the skills that you need," he said. "A clever network person is a clever network person. If you understand how a network is designed and architected, that's an amazing set of skills that will still be useful."

Looking ahead, Fox predicts the industry is nearing a tipping point.

"I think we're going to get to the truly self-driving network, and I think we're close at the moment," he said. "We're getting to the point where the network will be like a utility. You set it, connect it, and it will just run itself. You don't look at your phone every day and figure out what needs changing and configuring. You just use it."

"I predict that point where it is truly self-driving is a lot closer than we think."