Latest / Technology Now / Self-driving, intelligent, and built for AI: the future of networking in the 2020s | Rami Rahim
Transcript
- MICHAEL BIRD Well, hello, and a very warm welcome to a very special episode of Technology Now. The first ever recorded in front of a live audience. And we're here at HPE Discover Las Vegas 2026. I'm Michael Bird, and today we are going to be diving into self-driving networks and trust and asking the question, can we, as users, organizations, and technology providers truly trust a self-driving network? So to help answer that question, I am excited to be able to welcome to the stage Executive Vice President, President and General Manager of HPE networking, Rami Rahim. MICHAEL BIRD Welcome to Technology Now. Thank you so much for coming back. So what if we could maybe start off with the basics? When we say self-driving network, what do we actually mean by that? RAMI RAHIM I think it's sort of is exactly what it sounds like. You know, it's a network that takes care of itself. It configures itself, it optimizes itself. When there is an issue, it heals itself. The obvious analogy here would be a self-driving car, which we've, I hope, all experienced by now. I mean, there are literally self-driving cars that are driving around cities like San Francisco without drivers. They used to be a real novelty. Now nobody even takes a second glance at them. They're become so normal. The same sort of big revolution is happening to it, and it's starting with networking. And it just so happens that in the self-driving networking revolution, HP networks is a leader. MICHAEL BIRD How exactly do they work? Give me some details. RAMI RAHIM So good question. The way they work first is that you need to have a foundation of data, the networking elements, all of those elements that are providing the critical connectivity between users, between agents, between devices and things need to provide information that is fed into an intelligent agent. Basically, we call that Marvis. That's the AI engine behind our self-driving engine. That data has to be very granular and very specific. And to sort of summarize it, the thing that we want to glean from that data more than anything else is not whether the network is actually up and moving packets around, it is around the satisfaction of those that are actually on the network. That's the real unlock here. You don't want to know whether a network element, like an access point or a router is actually operational. Honestly, that's kind of table stakes. You want to know whether the people that are on the network are happy or not. Once you have that insight, the agent operates on it, learns from it, and it will start to address issues before they start to impair in any way the experience of those end users. That's what enables the self-driving technology to actually do its magic. MICHAEL BIRD Self-Driving. It's not an abstract concept like we have them in production environments. Customers have them in production environments, right? RAMI RAHIM The largest retailer in the world, the largest bank in the world, the largest health care system in the world. Many higher ed schools around the world have this in production as we speak today, and could not be happier. MICHAEL BIRD Has anything surprised you about how customers are responding or have been you adopting adapting to self-driving capabilities in their networks? RAMI RAHIM Any customer that hears this for the first time, at first has a healthy dose of skepticism because in some sense it is a little bit too good to be true. But I have to say, the most fun part of my job right now is to convert skeptics into believers. And it is possible, and I see it being done every single day. And the approach that we use is pretty simple. I can talk about it all day long. I can stand up on a big stage like the one I did a couple days ago at HPE Discover and talk about the merits that might hopefully get people intrigued enough to give it a shot. And that's what we're asking for, for customers to actually put it into production, put it on a floor, put it in a building, put it in a store and see it in action. And the action has real, tangible and almost immediate benefits after that, that tends to become the seed that starts to grow into more and more deployments, more and more layers of the network. And I like I said, I see this all day long, pretty much every day of the week. MICHAEL BIRD Why are we having this conversation now in, you know, mid 2026? RAMI RAHIM Why are we having it? MICHAEL BIRD Why having it right now? Yeah RAMI RAHIM Self-Driving networks have always been important. It's been something that we've been working on and perfecting and fine tuning now for many years. But what has changed today is that I believe we're moving for a paradigm where self-driving is sort of a nice to have, where one where it is actually essential. And the big thing that's different now is, of course, the era of AI that we are in the midst of. It used to be that best effort. Networking was good enough. It's just not good enough anymore. Right now, if enterprises want to become a genetic enterprises, tap into the latest AI tools to achieve greater levels of innovation and productivity in their organizations. You have to start with a solid foundation. And that's all the foundation starts with it, and the network and the network simply cannot be flaky. Show congestion, be slow. It can't slow you down. And as a result, I think customers today are looking for new tools that will help ensure that the investments that they're going to make in the application layers are going to be supported by those that they do in it. And so self-driving networks have become kind of the only answer to all of these sort of projects that enterprises are putting into place today MICHAEL BIRD when comparing it to a manual network. But yes, non self-driving, non self-driving words as a comparison. What does a self-driving network enable? That would be maybe difficult or impossible with a manual stick shift network. RAMI RAHIM Yeah. By the way, I do think there are some things in the world that should never be self-driving, like formula one cars. You know, like they should be manually operated forever, but networks should all be self-driving, and I think they will eventually all be self-driving. If I compare it, there will be some immediate benefits. First IT teams. I consider IT teams heroes because they are the ones that are basically, they've got their phone strapped to their belts and they are getting the calls at midnight, and they are required to make sure that connectivity is always happening at the most critical times for their organizations. And it's a thankless job, quite frankly. And typically IT teams spend like 70% of their time just troubleshooting firefighting. They will feel an immediate benefit where much of their time troubleshooting will be freed up to do far more important things. But it doesn't stop there. The more important unlock are the end users. You're going to see a dramatic increase in the satisfaction of end users as they use the network to do practically anything and everything today. I mean, naming an application today that does not depend on the network, it's practically impossible. You need to have an assured and secure network that can keep up with the demands of those users, and you want to make sure that those users are actually happy all the time. MICHAEL BIRD So why doesn't everyone have a self-driving network at the moment? Like, what are the what are the hurdles? RAMI RAHIM It's a good question. Everyone should have a self-driving network. I mean, the hurdles, quite frankly, are around trusting it first, trusting that it's actually possible because you haven't actually had your hands on it yet. You haven't seen it. But even once you see it, I think, and you see that the technology's there, it does take people a bit of time to build the trust in it. I, you know, going back to the self-driving car analogy, we didn't go from completely stick shift manual cars to completely driverless cars overnight. We sort of went to, you know, power steering and then cruise control and then lane assist and then, you know, keep your hands on the wheel, but the car will sort of drive itself to finally, we don't even need the driver. It was an evolution. The technology evolved in time. But then the trust level also evolved in time. And I believe this as similar sort of evolution or maybe even a revolution in the technology and in the trust of the technology is happening and networking right now. MICHAEL BIRD And when you speak to customers or people interested in exploring self-driving networks, what are the biggest concerns that people raise when they first hear about this self-driving network concepts? RAMI RAHIM I think the biggest concerns are moving from a paradigm where the network identifies issues and informs you of those issues, typically through red alerts or things of that nature, and then helps you diagnose them to a network that actually starts to remediate all on its own and just tells you after the fact. That is a kind of a pretty big trust leap. But the way that we've overcome that in our self-driving network technology is by giving customers ultimate control in capability, by capability, with literally toggle switches that allow you to say, for these types of issues, I will trust the technology to just go ahead and remediate and tell me after the fact. And in time, what we're finding is more and more of those toggle switches are flipped on, and eventually I think they're all going to be flipped on by default. MICHAEL BIRD I guess it's a control thing. I remember when I switched, like cruise control on my car in the first time, and I remember being like, oh, as it, like, approached the car for the first time. And then now I just, you know, put my feet up. You know, I don't really do that. RAMI RAHIM And that's just cruise control. So imagine getting into a car and telling it to just go to your destination and then like just, you know, waking up a few minutes later to, to be there. I mean, that's a bigger leap, right? MICHAEL BIRD But it's a control thing. It's it's like, you know, once I've done it a few times, it's fine. It becomes normal. I mean, I've not been in a self-driving car yet. I think if I went into a self-driving car, I think I'd be slightly freaked out by imagine there are people that ride them every day that it's the most boring thing ever. RAMI RAHIM Absolutely. I think the first time for everybody, it's going to be surreal, somewhat unnerving. But in time, everybody gets used to it. MICHAEL BIRD How is it possible for network, for a self-driving network to demonstrate that it's making the right decisions? RAMI RAHIM It's important for the intelligence behind the self-driving network to be as transparent as possible. Right. So we don't want this to be just a black box. We actually want to provide you with the insight that that the engine has gleaned from what's happening in the network to even provide you with the root cause, the different remediation that were considered, and then ultimately to provide you with an explanation of why the specific remediation that was selected was, in fact selected. I find that the more transparency and information that engine provides, the more trust is built in time. But but again, I want to be clear about this. I believe that these are all steps to eventually humans just trusting the technology. MICHAEL BIRD So it's sort of a stepped process to progression. You provided them with full control to begin with, but sort of encouraging them over time to be like, we got this. We got this RAMI RAHIM correct. Absolutely. Right. MICHAEL BIRD Okay. Can we talk security? Because security is pretty fundamental to trust. How is it possible that a self-driving network can stay secure? RAMI RAHIM So the one thing that we need to understand about security is that it's no longer a standalone domain. I mean, security has to be embedded into the network. And let me explain why every malware attack has one thing in common. It leverages the network to do its dirty work, to steal, to do damage, to do all the bad things that security tax do. For the same reason, we must leverage the network to create a more effective security posture. And what that means more practically, is that the network has to be the global sensor detecting attacks where they occur. And then just importantly, you want the network to be the policy enforcer, where you're blocking attacks, where it makes sense to block them. Right. So this notion of just having a few select firewalls and a perimeter is just disappearing. Unfortunately, that just does not work anymore. People are everywhere. But the thing that is always true is that people are connected to the network, things are connected to the network, and so the network must be used as your security fabric. So once you've got a self-driving network, you've got the intelligence that actually understands where there might be issues like cable issues or misconfigurations or so forth, and you can fix that. It's not a big leap to start to think, well, that same engine can start to understand what normal looks like for a network and to understand also when there are anomalies that look out of place or look sort of suspicious and as a result of that, become really, really good at detecting when there's something that's actually malicious, that's bad, and to actually put in place the blocks, the enforcers, to stop it from happening, or at least from spreading laterally from user to user department from department. MICHAEL BIRD And I mean, where do guardrails fit into this? Because organizations, I presume, will also maybe want to maintain some control over what the network can and can't do. RAMI RAHIM The control piece is universal. It applies to self-driving networks as applied to self protecting networks. So completely agree. It's important again to provide the transparency and also to provide the controls for the users to set the level of self-driving and self protecting capability that they want and they are comfortable with. MICHAEL BIRD And where does sort of security, governance and automation, where do they fit into a self-driving network? RAMI RAHIM Networks connect. People connect things. What you want is the security embedded throughout, and then you want to give the operators the ability to very easily put in place your intent, i.e. what you want to protect yourself from, the segmentation that determines what people have access to. Ultimately, the end state here is a universal zero trust network access. What that literally means is by default, you don't trust anything or anyone, and you enable certain people or certain devices to access certain applications on a case by case basis. The problem with this is that it's complicated, but with AI and self-driving capabilities, you can reduce the complexity dramatically. MICHAEL BIRD Apologies for the business speak, but I mean, is this going to be a bit of a paradigm shift in the way we think about our security posture, network security, RAMI RAHIM absolutely. Anything that leverages self-driving capabilities, whether it be around remediation of issues that might impact user experience, or it might be around just the simplicity of deploying policies for security, is going to be dramatically less complicated. MICHAEL BIRD Can we talk humans? Where do humans fit into a self-driving network? Are they completely excluded? Somewhere in the middle? RAMI RAHIM No, they're not completely excluded. But I think if you ask a network professional or an IT professional that spending 80 to 90% of his or her time troubleshooting, firefighting, diagnosing issues, hunting down misconfigurations, bad optimizations and wireless frequencies, and you ask them, how much do you like your job? I have a feeling you're going to have very few who are going to say they like their job. This is not the fun part, nor is it the important part of what humans do in it. But unfortunately it is the bulk of what humans do in it today. What we need the humans to do is to provide the intent. That's basically telling the network, telling the IT infrastructure to do what the business requires, what the strategy requires, and to focus on what's important, which is innovation, which is advancing the strategy of their organizations, which is to avoid being disrupted by smaller upstarts in their industries, to enable them to be the disruptors in their industry. That requires a strategic mindset to it, not a very tactical firefighting mindset to it. So humans remain extremely important, in fact, even more important. But their role elevates strategically in the organization MICHAEL BIRD so for IT professionals. It's about letting them do more of the fun stuff. RAMI RAHIM Absolutely. Yeah, fun stuff. And the more important stuff. MICHAEL BIRD So how do you balance the efficiency of automation with that sort of confidence and accountability that we get from human beings? RAMI RAHIM You know, in terms of efficiency, there are metrics that you can measure that we see coming from real world deployments. We have customers that have seen the number of trouble tickets in their organizations plummet by over 90%. So imagine that you as an IT. And we're dealing with hundreds of trouble tickets every year. And now you're dealing with single digits. What does that mean to the amount of time you have now to do the more important things? It's a complete, you know, 180 degree turn in terms of productivity for an IT team. And the accountability, I think remains with the IT team. Like nobody is going to blame the AI engine, you know, or assign the AI engine to full accountability. Ultimately, IT teams will remain accountable, but their accountability is going to be grounded in the trust that they put in the technology to do this firefighting, so that the IT teams can be accountable for more important things MICHAEL BIRD Okay, so for for organizations who are interested in in getting started, where on earth do they begin? RAMI RAHIM I think the best way to learn is to do. Okay, so you can go to seminars. There are probably all sorts of resources and educational capabilities that we provide as a company, but the best way to do it is to actually deploy it and run it. And you can start really small and grow from there. I promise you, once you see it, you simply cannot unsee it. And that's why typically some of the largest deployments in the world start with a tiny seed, where typically the IT teams themselves, see the unlocks, see the benefits, and then just want to expand it to everything. MICHAEL BIRD Okay. My final question was going to be about do you think all networks be self-driving? But you basically answered that at the top. So let me change a slightly. What percentage of all networks do you think should be self-driving? RAMI RAHIM 100%. It's just a matter of time. 100%. The only question really is, is it going to take one year or two years or three years? I mean, it's not going to be in the weeks because like I said, there's a technology aspect. We're not yet at a point where every issue can, in fact be diagnosed and corrected without human involvement, where maybe at like 80%. But I think we will be at 100% within the next couple of years. And then you've got to overcome the trust factor. But like it will be 100% at some point in time. MICHAEL BIRD Fantastic answer. Rami, thank you so much for joining us today. It's been an absolutely fascinating conversation. SAM JARRELL Okay that brings us to the end of Technology Now for this week. Thank you to our guest, Rami Rahim And thank you to our audience here in Las Vegas, and of course, our listeners at home. Thank you so much for joining us. MICHAEL BIRD If you’ve enjoyed this episode, please do let us know – rate and review us wherever you listen to episodes and if you want to get in contact with us, send us an email to technology now AT hpe.com and don’t forget to subscribe so you can listen first every week. Technology Now is hosted by me, Michael Bird, and this episode was produced by Harry Lampert and Eva Higginbotham with production support from Alysha Kempson-Taylor, Beckie Bird, Alissa Mitry, Katie Champagne, Zoe Revis and Jenessa Ayache. Our onsite production team was Alvaro Calero, Molly Zapollo, Serry Fox, Karina Fletcher, Kimberly Evans, Luke Smith, Brent Burg, Chase Jensen, Kevin Fritz, Eric Zeisger, Erica Cornwall, Andre Hins, Josip Andric and Robert “Doc” Burt, big team. Our video editor was Leon Radschinski-Gorman and our theme music, as always, was composed by Greg Hooper. SAM JARRELL Our social editorial team is Rebecca Wissinger, Judy-Anne Goldman and Jacqueline Green and our social media designers are Alejandra Garcia, and Ambar Maldonado. MICHAEL BIRD Technology Now is a Fresh Air Production for Hewlett Packard Enterprise. (and) we’ll see you next week. Cheers!