- Network vendors like Cisco, Extreme Networks and HPE Aruba are automating systems with traditional machine learning and new LLMs
- AI can turn vast amounts of network data into actionable insights, but full automation will take time
- The goal is eventually to integrate AI into all aspects of network operations for more intelligent, adaptive systems
The network has become the central nervous system of every business, housing and transporting vast amounts of data. The trick is finding a way to turn that data into actionable insights. That's where artificial intelligence (AI) comes in – the tech is helping network vendors steadily automate the systems that keep businesses connected to prepare for a future in which the networks might one day run themselves.
"There's a tremendous amount of data and information just sitting out there in the network," said Dan DeBacker, SVP of Product Management at Extreme Networks. "The ability to quickly and easily harness all of that data and turn it into usable information is where the potential of AI lies."
Questions have arisen about AI’s killer use case. But rather than a single breakthrough, DeBacker said AI's value in networking is more likely to come from a collection of small, incremental improvements.
“It's a little thing here, little thing here, little thing here, and all sudden, you’ve got 100 little things that build up into something really, really big,” he told Fierce Network.
Despite what the hype may lead you to believe, AI is not seen as a standalone product, DeBacker said, but as “a capability that will be integrated into everything we do.” Eventually, as the technology matures, the network will become more intelligent, capable of adapting to new challenges with minimal human intervention.
For Extreme Networks, this means continually adding more advanced automation and AI across its stack. "We've been doing automation forever, right? 20-plus years," DeBacker said. "But as we continue to mature these capabilities, continue to train the models, and the data continues to get better and more refined, the automation becomes much more intelligent."
Automation is the goal. 'We're not there yet.'
AI's impact isn't limited to network management; it's also reshaping network security.
As networks grow more complex and prone to attacks, overstretched IT teams are finding it difficult to maintain performance and security. "It is becoming increasingly challenging to prevent service degradation, remediate attacks and deliver consistent, high-quality digital experiences to a hybrid workforce," said Matthew Landry, VP of product management for Cisco Wireless in a recent op-ed for CIO.
To simplify and streamline network operations and reduce the burden on IT teams, many decision makers are considering network operations solutions that are AI-enabled, added Landry. These includes tools that use traditional machine learning (ML) and others that are beginning to tap into large language models (LLMs) to make it even easier to draw information out of the network.
A benefit of AI is it “can see and understand things that are beyond what a network administrator or a networking team can," explained Larry Lunetta, HPE Aruba's VP of AI, security and networking product marketing. "Networks are just too complex. You can't rely on manual processes anymore to manage networks or support to secure them. So that's why AI is increasingly important.”
While AI can automate many tasks, full automation is a gradual process that requires building trust in the technology, Lunetta noted. Humans are still in the driver’s seat since AI models require oversight.
"The forward-looking vision for AI is to get to the point where automation is feasible and customers will accept it," Lunetta told Fierce. "We're not there yet, but we're continuing to build that, and the more experience we get with these solutions, the more likely we can get to that point. But automation is clearly the goal."
Thus, the total takeover of AI in network management and security may still be on the horizon, but the groundwork is being laid, one tool at a time.