Innovations with Security-First, AI-Powered Networking - Six Five In The Booth at HPE Discover Las Vegas 2025

How is networking evolving into a truly intelligent, self-managing, and security-first fabric for modern enterprises?

Discover the major advancements from HPE Discover Las Vegas 2025! Join Will Townsend as he hosts David Hughes, SVP and Chief Product Officer, HPE Aruba Networking, for a conversation on the innovative strides that HPE Aruba Networking is making in optimizing networks through a focus on security-first, AI-powered technologies that redefine IT operations.

Key Takeaways include:

‍

🔹Agentic AI for Hybrid Cloud Operations: An in-depth look at GreenLake Intelligence, an industry-leading agentic-AI-powered hybrid cloud solution, and its revolutionary impact on IT operations, particularly how it's implemented within HPE Aruba Networking Central to transform network management with agentic mesh and networking copilot capabilities.

‍

🔹Real-time Network Orchestration: Explore the advanced functionality of agentic AI within HPE Aruba Networking Central, detailing its evolution beyond monitoring to encompass real-time orchestration and copilot capabilities for precise problem identification, root cause analysis, and automated remediation.

‍

🔹Tangible Customer Benefits from AI: Practical examples of how these cutting-edge AI advancements are delivering measurable benefits to customers, such as significantly reducing latency for critical applications like Slack through proactive and intelligent network adjustments.

‍

🔹Scaling AI with a Robust Data Foundation: Insights into the continuous development and expansion of AI features within HPE Aruba Networking's platform, shedding light on the exponential growth of their data lake and how this evolution directly supports increasingly advanced and efficient network operations.

‍

Learn more at HPE.

‍

Watch the full video at Six Five Media, and be sure to subscribe to our YouTube channel, so you never miss an episode.

‍

IMPORTANT: Note to Uploader: VIDEO EMBED - added by uploader.

‍

Or listen to the audio here:

Disclaimer: Six Five In The Booth is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

‍

Will Townsend: This is Six Five In the Booth at HPE Discover Las Vegas 2025. I'm Will Townsend, I manage the networking and security practices. For more insights and strategy, I'm joined by David Hughes. David leads overall product for the HPE Aruba business unit. David, it's great to be chatting today.

‍

David Hughes: Yeah, no, really good to be here.

‍

Will Townsend: Awesome. I loved your keynote this morning, but I gotta tell you, I'm sad that Dan is no longer with us.

‍

David Hughes: Oh yeah, well, we had a lot to pack in there and something had to go.

‍

Will Townsend: I hear you. The demos were super insightful, I learned a lot and congratulations and there are a lot of announcements to talk about, but I want to start with the agentic AI Mesh. And this is Aruba Networking, basically both generative and agentic AI offering. Can you talk a little bit about that?

‍

David Hughes: Yeah, no, it's really, really exciting. So what we're doing is applying agentic AI to the networking operator's job. And if you think about it, what agentic AI does is it gives the IT team some extra IT members that are working day and day out while they're sleeping.

‍

Will Townsend: Yeah. Antonio said during the keynote, like while you're sleeping, while you're eating, they're working behind the scenes, right?

‍

David Hughes: Absolutely. So in the architecture there's a Copilot that most people are familiar with which is kind of prompted by the admin. Right, so the admin types of your copilot answers. What's really unique about agentic AI is we have a separate LLM, another very large LLM where the reasoning and action capabilities and it's headless, it's prompted by things like network events, like maybe it detects some kind of anomaly and then that triggers a root cause analysis and then any remediation actions and that reasoning agent is able to access Vera mesh. Many, many sub agents. Each of those sub agents specializes in a different thing like maybe link health or identity information or documentation or bug records. And the agentic multi agent orchestrator is able to ask all those questions, correlate all that information and then assemble a chain of thought like this is what happened, this is the steps I took to understand what the root cause was. So this is the root cause. Here's two or three ways you could address it, the pros and cons of each and here's what I recommend you do. And then it gives it back to the co-pilot who presents it to the network operator who can decide, yes, this makes sense. I trust AI. I'm just going to say yes or maybe like many people, they want to double check so they can look through that line of reasoning and say, yeah, this makes sense, I'm going to go with it.

‍

Will Townsend: Sure.

‍

David Hughes: So I think it's going to be, to me, it's bigger than Generative. Generative gave us a nice natural language interface last year. This really augments the IT team with extra resources that go run things down. It would take hours otherwise and then assemble it in a way like if you hired a new grad admin, you give them a job, they come to you. Here's what I discovered. You're going to read it and then say, yes, no, it's like that. But you're getting it, as we said, while you're sleeping. So I think it's a real game changer.

‍

Will Townsend: I believe it is too. It's going to drive automation at scale. It's great. Going to work very nicely with your opsramp integration as well. And you talked about that today, the observability of not only HPE infrastructure, but multi vendor infrastructure as well. Right. But my understanding is that what you're doing with the agentic AI mesh is part of sort of a larger vision that the company has around Greenlake Intelligence. And so can you talk a little bit about that and how the two fit together?

‍

David Hughes: Absolutely. They're almost one and the same. So what we've demonstrated is kind of using this agentic AI mesh within the networking domain. But our vision is to do this across HPE and that's what we call Greenlake Intelligence. So imagine that that copilot and that multi agent Orchestrator are running at the HPE level. And as well as talking to the sub agents I discussed, they're thinking about our multi agent Orchestrator as one sub agent of that agent and a storage agent and a compute agent, a workload based agent, each of which then may have more sub agents. But you've got all of this information, all these agents working together in a mesh and then you're able to kind of answer and troubleshoot across domains. I think back to the multi vendor piece. I think what's really key there, especially with opsramp, is being able to get the telemetry in from all different vendors and across domains. Otherwise, even if you're doing a good job, you're only looking at a slice of the network or a slice of the problem. So of course I'm very excited about it inside the networking domain as my business, but it's going to have a big effect as we also bring it to life across HPE and the whole.

‍

Will Townsend: Portfolio, I believe it will too. And again, what it's going to do, it's going to reduce those gaps in visibility. Right. And it's going to improve network assurance as a result and it's going to improve security posture. And I know that you're very focused on a security first architecture and I'm going to go a little off script. But you talked about zero trust today as well. Can you talk a little bit about that and what you're doing to improve zero trust?

‍

David Hughes: Yeah. For us, Zero trust, there's sort of two very important technologies. One is the SSE and SASE piece where you're providing policy that applies to users and users identities and applications are allowed to access and then potentially with things like CASB, what they're able to do with that application. Sure. So we've invested there and we're on a path towards tighter and tighter integration. The other piece is what do you do about devices? Because device identity is different from human identity and they're accessing services and data versus existing applications, the IoT devices. And so what we're doing is evolving what we've done with network admission control where we're able to identify different classes of IoT devices, continuously monitor them, track their behavior and then apply policy which can be implemented on prem versus in the cloud. And so as we bring that kind of zero trust for devices and zero trust for people together, that's what we call universal CTNA and it's an area that's of particular focus for us.

‍

Will Townsend: Yeah, I mean it's powerful infusing networking and security together. I think today you talked about it's not a bolt on because when you bolt on a security solution it creates gaps. Right. And so I really like what the company is doing there. Now I know that what you've announced this week is in customer preview and I believe it's going to be generally available in the third quarter this year. But can you speak to some of the applications and use cases for AI Mesh?

‍

David Hughes: Yeah, well, it's super flexible because it's very, very powerful generic technology. But two examples. One is the example I showed where there's a problem where the users are experiencing some latency when they roam. That's the problem that's actually determined using our classic AI. Traditional AI figures as an anomaly in the performance. That's what triggers that multi agent orchestrator to go off and start digging into what's going on and it realizes that hey, there's a problem with this class of device and this class of device based on our bug database is incompatible with 802.11. I'm getting technical on you, but that's the thing. It can dig into all this stuff and figure out that we should maybe turn that off. But this would be the cons. But on balance it makes sense. So there, that example is going really, really deep on wireless troubleshooting and remediation. But the same technology is equally able to drill down to a switching problem or SD WAN problem. And another example that I think you might see on the booth over there is there's some users that are experiencing a poor experience with an app and that's what triggers it. And the root cause in that case is that on the WAN gateway not enough bandwidth had been assigned to that particular application class and so the recommendation was to increase it. So it's really, really very powerful. And one of the things I think that really excites me is that people have been afraid of AI, generative AI in particular, because they don't understand how it gets its answer. It's not explainable. Agentic AI gives you that explanation. And what's really cool is I really liked it the way we embed graphs and everything in that flow. So in a user interface oftentimes people want to do all different things with that. So you end up with a lot of controls so that say someone can get a chart of anything and you end up kind of reinventing burst or whatever inside your new life.

‍

Will Townsend: It's really kludgy and hard to navigate.

‍

David Hughes: But with AI and generative AI even you can describe the chart you want and you can kind of generate it on the fly without having to have it fit into a part of a ui. It can kind of be more in a workflow, in a document type flow. So that's what we're seeing there with that reasoning agent generating a workbook of all the things it's done.

Will Townsend: So it's not behind the scenes anymore, it's front and center, right?

‍

David Hughes: Yeah. It's cool.

‍

Will Townsend: Yeah. That's an interesting observation. Well, David, as we wind up our conversation, I want to talk about the importance of data, data feeds, these large language models and small language models. And HPE has a tremendous data lake. Can you spend a little bit of time talking about that and the advantage that you bring with your solution? Because there are competing solutions out there.

‍

David Hughes: Yeah. So with Aruba Central now we're managing about 6 million devices, that's APs and switches, and then attached to them are about 3 billion billion endpoints. So it's phones and laptops and IoT.

‍

Will Townsend: Pretty mind blowing when you think about the number.

‍

David Hughes: And then for those, every minute we're collecting telemetry, telemetry relative to the operation of the networking device, but also the experience of the people on those clients. So more than a trillion data points a day, which we stream in through central, we anonymize that and place it in a data lake. And then that data lake is really the foundation that we use to learn against, to be able to kind of do fleet learning across all our customers, have the AI identify by itself what are best practices, what are poor practices, how can learnings overall from their fleet and apply them to individual customers to give them a better experience. Or if you're trying to identify IoT devices, you're able to do it across this huge data set. It's a much easier problem. So as I said in my chat, telemetry is really the lifeblood of AIOps. And that's also why we're so excited to be able to partner with opsramp and bring the third party telemetry in. Because the third party telemetry now goes into the data lake and it gives us an ability, it enriches everything. Right, Exactly. We can send a full topology versus just a slice of it.

‍

Will Townsend: Yeah, yeah. Well, you know, I followed the company for many, many years. I knew Kirti and Partha many, many years ago. We met when you joined as well. And you're very credible in this space from my perspective because you've been a leader in AIOps. When you look at your edge services platform dating back a couple years ago and just what you've done continuously with machine learning and with AI, I think the future is very bright, David. So I want to thank you for taking the time to chat with me. It's been great. And I want to thank our viewers for tuning in. This is Six Five In the Booth at HPE Discover 2025. There's lots of more content out there, so stay tuned.

‍