Hewlett Packard Enterprise (HPE) this week unfurled an updated set of frameworks that leverages artificial intelligence (AI) agents running on microservices-based platforms to automate network management.
Mittal Parekh, senior director for product and solutions marketing for HPE, said the core agentic AI capability now makes it possible to deploy “self-driving” networks that can be centrally managed via the HPE Aruba Central console or the HPE Mist console the company gained with the acquisition of Juniper Networks last year.
Specifically, IT teams can now rely on AI agents to autonomously identify capacity bottlenecks and dynamically tune wireless network parameters such as band selection, channel bandwidth and power levels.
Additionally, an AI agent can be employed to autonomously fix VLAN configuration errors or detect and remediate unauthorized DHCP servers.
AI agents can also be used to ensure smooth, uninterrupted roaming for users by analyzing client connectivity metrics, including location, and surface the root‑cause of latency issues. An AI-driven Radio Resource Management (RRM) capability can also adaptively learn to avoid association issues on frequently impacted channels.
Other capabilities added to the HPE network management platforms include simplified inline microsegmentation to enforce security policies and a network access controller (NAC) sandbox testing tool that enables policies to be validated against actual network conditions.
Overall, the goal is to eliminate the need to rely on human network operators to perform low-level tasks, said Parekh. In the age of AI, mean-time to resolution (MTTR) will now be achieved in seconds, he added.
In fact, as AI agents continue to develop more reasoning capabilities, networking as it was once known is in the past, said Parekh. The goal now is to take humans out of the network management loop altogether, he noted.
At this juncture there is little doubt that AI will be relied on more to automate a wide range of networking tasks that today take too much time for largely understaffed IT teams to consistently perform.
Tom Hollingsworth, an industry analyst with the Futurum Group, said HPE’s self-driving network is squarely aimed at IT teams overwhelmed by operational frustrations. Proactive problem resolution is critical to stay ahead of the challenges that are faced by operations teams that are today typically run lean, he added.
Regardless of the approach to applying AI to IT operations (AIOps), the challenge is, as always, funding the infrastructure upgrades needed. HPE has made available two financing programs to provide a 10% savings on data center networking and enterprise routing for AI workloads, and 0% financing on HPE Networking software. However, given all the competing AI projects being launched within an organization it may be difficult to convince business and IT leaders to prioritize investments in next-generation network infrastructure.
Eventually, however, the day will come when the overall size of the IT environment that needs to be connected exceeds the ability of existing teams to manage. At this point, automating network management will become much more of a necessity if organizations hope to be able to run AI applications across IT platforms and endpoints that are only going to continue to become more distributed with each passing day.
