Modern Networking for AI vs AI for ModernNetworking
34m 25s
This podcast episode discusses key announcements from Cisco Live 2025, focusing on new Cisco networking switches like the 9350s and 9600 series, which incorporate embedded DPUs and advanced AI functionalities. The conversation highlights a shift from legacy, manual network management using command-line interfaces to AI-driven, proactive operations. Experts explain the difference between "AI for networking"—using AI tools to improve network engineering efficiency—and "networking for AI"—designing infrastructure to support AI workloads, such as GPU communication. The Cisco Meraki dashboard, integrated with AI agents, is emphasized as a user-friendly platform for deploying these technologies, enabling automated troubleshooting, predictive analytics, and enhanced security. Use cases include automated patching, threat isolation, and performance monitoring, helping organizations scale securely while mitigating risks like cyberattacks or bandwidth issues. The episode underscores the growing demand for modernized networks to meet evolving business needs, with Cisco's solutions offering both hardware innovation and software-driven intelligence.
[MUSIC] Digital transformation has become one of the busiest of bus words. This podcast is where we decode and deconstruct digital technologies, strategies, and best practices that are impacting the way companies deliver value to their customers. No bus words, no BS, let's get to it. [MUSIC] Hey listeners, it's your host, Alec. And before we jump into the podcast, I have a quick plug that I think is an awesome opportunity to expand your cloud knowledge. I highly recommend checking out Procidio Explores, which we will have linked in the show notes. Procidio Explores is your virtual adventure that will take you through a number of journeys that will have you scaling new heights in cloud collaboration. Thank you, and let's get to the episode. Hello everyone and welcome back to another episode of the Digital Decode brought to you by Procidio and our partners at Cisco. I'm your host, Alec Brust. I'm actually going to jump right into it because we are actually talking a little bit about our partners over at Cisco today. And the reason we're talking about our partners at Cisco is that at Cisco Live this year, 2025, Cisco announced several new switches. And I'm not going to dive into it because I have my experts on with me from the Procidio side to talk about it today. With me, I have Andy Richter, Technical Director at Procidio, Mark Cardwell, Director of Professional Services, and Brittany Bolton. Solutions architect, how are you all doing today? I'm doing fantastic, Alec. It's great to see you. Yeah, it's great to have you all on. Andy and Brittany, you've been on with me before, but this is Mark's first time, so welcome Mark to the podcast. We're really excited to have you. Thank you so much. Good to be here. Of course, great to have you. Andy, you just heard that little intro I gave you, but I definitely want to hand it over to you to just kind of give us, you know, what's the rundown of what happened this year at Cisco Live. Thanks, Alec. It was a big year here for us at Procidio at Cisco Live, and there was a lot of really exciting announcements that came out of networking with Cisco. And a lot of new platforms on the Cisco announced the 9350s announcements, the 9600 and new data center switches, the N9300s with embedded DPUs. And let me tell you, the thing that was just so interesting is that, you know, Cisco has always been a company that's had impressive hardware. And Cisco's hardware, this year is impressive, just like all the other generations have been. But what's been different this year is just a really wonderful evolution. And in some case, revolution in how we can do networking with using AI, using these modern functions, and really Cisco helping us lead the way in improving network operations. I've just been having so much fun talking with friends here, more converting about what these new platforms can really bring, you know, not just for us here at Procidio as we help our clients, but also, you know, how our clients use their networks to gain more business value. So I think it was a big year here, I guess, Alec is the answer to the question, and we couldn't be more excited. No, that's great. That's great that you're excited about this. I wish I shared the same excitement and after talking to you guys a little bit through our outline sessions and things like that, I definitely am starting to. And so I think, thank you so much for the introduction. First of all, Andy, but I mean, I kind of want to talk about this evolution. So what I think we'll start with Alec is, I want to talk about, you know, Brittany, you know, you've had a lot of experience here as a Procidio architect. You know, I've known you professionally for a long time now, and what our listeners probably don't know is you've been an architect, an engineer, a project manager, and you've seen projects and networks through all kinds of different generations of this technology. And I know you've seen the old call legacy technology legacy projects, you put up post and personal with me seeing some of these new ones. When you think about that, like, the old school data life of how we use the troubleshoot things and the old way of doing stuff and you see some of these new AI, agents, like, operations, giving it any of the kinds of things you think about having to do in the past. Maybe in the future, you'll look forward to doing less of. So it's funny to hear you say that like, I don't feel like I'm old enough to be, you know, associated with legacy equipment, but we're just going to skip right over that. And I'm going to say something that might be slightly unpopular for a minute, and we're just going to get it out of the way. There is a part of me that is nostalgic for the days of old Andy, like, there was something just so cool and exhilarating about being one of the elite who understood the command line interface. There was something really exciting about that. However, we also know that it wasn't exactly the most efficient way to go about troubleshooting an issue, right. I mean, even just think about the potential for human error when you had multiple tabs of command line open and you could accidentally type the wrong thing in the wrong window and then next thing, you know, you have a big problem on your hands, right. So, you know, your process was very different. We had to be not only meticulous and very careful with what window you were in, but something simple, such as, you know, configuring a VLAN throughout a network was very manual. You had to go to every top of rack switch, right. These things are changing and I don't think that's a bad thing as much as I have nostalgia sometimes. I think that's all for the best, right. Think about all of the other things we can spend time on when we have efficiency brought to us via AI. So, you know, when you talk about config errors, you know, I often feel personally seen by that Brittany and I can appreciate that because you know, we've spent a lot of time working hard at this and more, you know, I look over to you and I think about what our clients are looking for and why they're looking to us for these kinds of solutions. And I know you have a, you know, really a wonderful professional services team who does networking here for us for CIDIO. And I know that the demands on you and your team are, you know, always increasing. You're always getting more work and you're always looking to do bigger and better networks. How have these kinds of automated and agentic tools given you new and bigger opportunities to solve some of those old problems in bigger and faster and better ways. So, you know, you know, you know, kind of going where where Brittany was in the older nostalgic days, you know, we would spend time, diving through just loads of configurations trying to figure out where things may be problematic where we'd be running into issues trying to design a better network or a better mouse trap really, you know. As we move into more of a agentic style network ops and switches that actually will start to think and really through dashboards such as Maraki's dash, Cisco's Maraki's dashboard powered by an agent, AI. And then you're going to be able to sit there. It's going to be able to look at what's going on in the network and not react, but pro act go into and actually see what's going on. Look at data, telemetry data that's coming in and say, I've seen this before. We're going to be able to do that. We're going to be looking at that and go in, wow, I'm still about two hours behind. And, you know, whereas the power behind what Cisco is producing within their platforms will allow for that to be sped up and allow better for us to go in and get deeper into the technologies and be able to be with our customers providing them more. Of a outcome based solutions so that because what we're not doing is having to sit down and I got to get into the nitty gritty of trying to troubleshoot into this and it's going to allow for us really to to focus more on that customer and really provide the solutions that's going to allow them to go into that next realm of data delivery. You touched on so much that I want to click on so long that I'm going to pick on us a couple of highlights here mark at a high level I want to kind of work on talk about a eyes role in modern networking and vice versa the role of modern networking in AI. And so, you know, I talked to a lot of our clients about these kinds of outcomes that you were talking about predictive analytics using AI to help with diagnostics, you know, automation of routine tasks. These are things that, you know, are both aspirational for a lot of our clients and achievable for a lot of clients as well. But, you know, what I'll say you can do these things they say well, I don't have a giant GPU farm to pull from and I say well, you don't need one work. So, you know, when you talk to clients about what the differences between, you know, AI and networking AI, can you define kind of those two things and then I want to kind of come back really talk a little bit more about how you consume and how you use AI for networking. If that's okay with you, sure. So, kind of AI and networking. So, part of that has to do with where these models are lying. So, a lot of customers may be thinking that well, I need to have large scale LLMs or large language models on my my network in order to take advantage of some of these these newer features and it really isn't that sure. You can, but utilizing the applications that are AI aware, thousand eyes, others along those lines really allow for that that AI to come into the network, but not be run into the network. Now, the new access, what is it, C9350s.
They're actually being designed to allow for if the customer wants to be able to run those GPUs, it's going to allow for that communication. So it's allowing that AI to be on the network. But it also allows it to be in that network. So allowing for a better ASICs and larger bandwidth within the switches to allow for the communication between those servers is definitely key. But if you look at really not having to have those and really having that the ability, whether it is through the dashboard or through a cloud-based enterprise class, software such as Thousand Eyes to be able to run and to utilize the AI applications or the AI models to be able to see those areas. So Bernie, when we look at all the things that mark just kind of went into there, there's so much that I totally agree with and I think it's so important. I want to look at this from just one lens to double click on. And that is there's a, we look at AI for networking or networking for AI. And I think when we talk a lot about you know, agentic operations and AI-powered diagnostics and AI Canvas, a lot of these really amazing features that Mark talks about. The first thing that I hear from a lot of my clients is, you know, well, I don't have a GPU farm to do all that stuff. And you just don't need all of it like that, Brittany. And so I'd like you to just kind of talk me through, you know, your thought process and how you work with clients, so to talk about what the offerings are, how things work, what the different components are, how networking and AI kind of work together in this kind of symbiotic relationship. But yeah, so Andy, that's a really good question. And I think it boils down to meeting customers where they are, right? So we know that the conversation around AI can seem extremely daunting and it can seem a little bit intimidating for, you know, customers that are maybe just dipping their toe in the water. So I think it's really important to really understand where they are, what their appetite is for growth in that space. And then kind of tailoring our approach accordingly, right? Like let's crawl before we walk before we run. And I realize that's maybe an overstated consulting process, but it's there for a reason. There's value in absolutely guiding them through the first steps, which may or may not be, you know, AI for networking versus networking for AI, right? Andy. And I think that distinction is a little bit fuzzy sometimes as well. And in my mind, and keep me honest here, Andy, but, you know, most people can utilize AI for the purpose of being more effective network engineers. And you don't have to have a GPU farm for this, right? You can use, you know, insert your flavor of preferred AI here and allow it to bring you more efficient use of your time in your role as a network engineer. So that's a possibility. And then we roll into networking for AI, which is a slightly different conversation. And to Mark's point, this is where it gets a little bit more complicated. And we have to look at things like, you know, potentially licensing, etc. So I want to get it right back into use cases here guys, because what we talked about, you know, Bernie, you know, networking for AI and AI for networking. And I do want to focus on how we can use AI for networking and how things are fundamentally overall integrated and functional. The thing that I just find so amazing, particularly about these new Cisco switch platforms that are coming out in the market here that were announced in Cisco Live is that it's pretty straightforward to get AI for networking running in your Cisco environments. And so, you know, my understanding of really the manual here, looking at the slides that came out in Cisco Live Bernie is that you can get these new Cisco switches, the 9350s, integrate them into your Cisco and Marockie dashboard and utilize a AI agent out of the box. Is that true, Bernie? And what do we think about it? That's my understanding as well. And I think it's really going to make, you know, AI for networking more approachable. That's what I think. Because when I think of, you know, ease of use and approachable networking, I typically do think of the Marockie dashboard, right? It brings you that visibility all in one place. It brings you that ease of configuration in one place. So it's only logical to me that this would be a place to start with the 9350s and bring in AI out of the box to the Marockie dashboard. And Andy and Brittany, can I ask you a quick question? And I'm not sure if this is maybe too early to tell, but have we been seeing a lot of people, you know, making the switch over the switch to these new switches and having these new switches in their networks? I've been seeing a lot of people adopting this or is it kind of slow to roll out? I'm going to route that one to Andy. So I will say that we have seen a lot of interest with every new Cisco switch platform. There is always a supply chain to get all these switches out the door. Absolutely. We have seen an incredible demand for switch modernization. Because we've seen a significant number of our clients who have had laggies, he switches their environments. And they've been doing things that we talked about, like, kind of the old way for a long time. Like Brittany was talking about the old way of doing things. And they're looking for a new way of doing things because their requirements for their network isn't getting any smaller. Like Marock was saying a few minutes ago, we're seeing so many increased demand for use cases like edge inferencing, micro segmentation, detailed network analysis all the way at the edge of the branch. But you can't just do it the old way anymore. When we look at this and our clients approach you, you know, what are the kinds of problems that you think about solving kind of at scale organizations using these kinds of tools, using the Cisco Marocky dashboard and AI chat thoughts. So a lot of it has to do with these basically ease of maintenance. And really, I mean, Brittany kind of hit on it just with just being able to come in and manage those devices. So you know, you've talked about automation of tasks, you know, and in the old days, we used to, you know, definitely run scripts. And Brittany was keying in, you know, well, I may, I'll run a script. Yeah, well, that's, helps in time. But what happens if I run that script on the wrong tab or so where, where we're looking to help our customers is one in that automation and task automation and better maintenance of their switches, being able to have clear databases of their configurations. And being able to, let's say, part of the C series, the accessories, the 9350s is really embedded security. So, you know, being able to help our customers, you know, forget day one, you know, day one your day late. So, and day zero, I think you're even late, even if you're trying to hit those day zeroes, you know, the new abilities of AI and insecurity is being able to actually it to go in and look and determine and see well before anything actually comes in or as it's coming in, detected immediately and shut it down so that, you know, zero, basically zero access, you know, just not being able to even get in the front door. The foot just gets slammed immediately. I think those are a lot of what we're seeing just with the attack vectors, the way that they're just broadened so much. And I mean, people using AI, well, now AI itself, using that as a tool, they come in. And so, if we're able to use the AI to defend against it, I think that's a lot of where I'm going to see and where we see the conversations going with our customers and being able to build that network that is really going to help our customers in shutting down, you know, even if it's not a a a taxa, but in bandwidth spikes where you're getting, you know, a talkative port that, you know, may be flooding your network. Well, if I have a huge pipe now and I have something that goes awry, I'm going to flood that pipe. It's even bigger. So now I that errant device can actually do a lot more damage. So being able to see that performance wise and using the dashboard to be able to mitigate those and really to come in and see those performance affecting incidences, let alone just this year attacks from threat actors. You know, Mark, I love listening to you talk about these kinds of tools and strategies because I hear you talk a lot about cybersecurity, you know, threat isolation, reacting to threats, but I also hear you talk about availability and, you know, the scalability impact on networks. And so, Alec, the fun part of listening to Mark talk is that he's talking about bad actors that can hurt us in different ways. They can hurt us by attacking us by knowing the specific exploit, the precise mechanism to get into that to a vulnerable software, or attacking a network by overwhelming it by it with too much traffic. And, you know, using AI, we can look for both of these types of threats.
And it's so amazing that even if we haven't seen this kind of specific attack before, the strategy's Marcus talking about there are available in these Cisco solutions to help us actually mitigate these threats before they're actually impacting our networks. >> Exactly, again, we're just going back to that proactivity versus reactivity. And I think that more touched on it perfectly. >> You know, Brittany, I want to double click on this a couple of things here. There are a couple of use cases that I heard from Mark that I think that I've heard from my clients. I wanted to just toss them over there and your court one is automated patching of solutions while not necessarily an AI feature, but a core component of any cloud managed platform. Is there kind of a world where we're going to be looking at using these AI features and functions with non-cloud managed infrastructures? That really kind of what we're going to be using is table stakes for networking these days. You kind of got me on that last phrasing about whether or not that's going to be table stakes. And I feel like that's a little bit yet to be determined. However, I want to go back to what you had said and this ties into what Alex said about proactive versus reactive, right? We know that in order to maintain a healthy network, we have to patch. And because of the threat vectors today and how things are changing to Mark's point, if you're doing it day one or even day zero, you're already behind the wall, right? So logically, we should be proactive about things like patching, right? And being able to isolate that suspicious activity, whether or not it's actually been proven that it is a bad actor is irrelevant. But the fact that we can view that suspicious activity proactively isolate that portion of network and avoid disruption is a really big deal in my mind. And I think that is the way of the future, right? And I think the other piece that's interesting is it's no longer a requirement to be sort of genius person to hack a network or to hack in general, right? And that's both interesting and terrifying at the same time. These people, young teenagers, for example, are able to utilize AI to build ransomware, to build effects. So why wouldn't we use AI to combat that very same thing? Yeah, absolutely. I couldn't agree more. Mark, this all adds up to me to say that this is a real inflection point when I look at networking and enterprise networking in particular, that when we've looked at the evolution of networking that in the business that we've been in, in the journey we've taken professionally is that, you know, we've gone from Catawest to iOS. We've seen the evolution of NXOS with all these amazing wonderful features that's just come out with. And we've seen this new generation of switches and routers and security tools with AI integrated in them. The demand of business are increasing so rapidly, you know, isn't really appropriate to just replace our infrastructure one for one, is doing it the old way good enough anymore? Or are these new technologies so important that we need to think about a new way of doing networking as we modernize our infrastructure? So if your network requirements are stagnant, sure, I mean, there's no reason to, but I can tell you right now, I know of no customer whose network environment is static. They may feel that it is, but it's not. Their requirements, their needs are just exponentially growing. And I would challenge them to say, look, if you don't feel that you need, if same, same is, is, you know, okay for you, how about there's things that we were just talking about? The fact that we have a tech services out there from the internet, your older switches or the same switches that being brought in that don't have these abilities into them for proactive management for proactive security, really zero trust out of the box, then you're going to be pretty much out of business because you're going to be shut down, you're going to have a ransomware attack, you're going to have some denial of service that's going to come in and really just shut your network down and your employees aren't going to be able to work. But also looking at the sheer fact that you don't want your business to stagnate, you want to grow your business. So that network is going to need to grow with you. So I may be looking at this and saying, okay, well, can I afford or am I able to do this? And the answer is yes, because the flexibility of these new platforms to be able to come in with licensing and be able to have it be one thing now, but just with the ability to add a license to then enable other features on top of that. To me, it's just a no brainer. I wanted to double click on that because you talk about unlocking additional features. And I have to tell you all, I am so excited about cyber vision. So I know that wasn't necessarily like the point of today's conversation, but just to add to that, right? When we think about like telemetry and network assurance, we obviously think of thousand eyes. We think of these new embedded security features. But can we talk about what cyber vision is going to do for the OT landscape, right? We are now bringing that same level of network visibility and assurance and proactive capability to the OT space, which so desperately needs it, right? So that's a shameless plug. I had to throw that out there and I'm very excited about that. And it is an additional license add on. So people who are looking at, you know, advantage already for thousand eyes, guess what? Cyber vision is part of that. So just throwing that out there, I'm excited. The hardest thing about my job here at Presidio Alex is creeping up with engineers like, like Bertie and Mark here because they're both so sharp. I want to start with where Mark was about the strategic advantage that businesses can see when they do this modernization, when they can achieve these advantages through more efficient connectivity to connect their devices together, but also protect their devices from these threats. As you said, Mark, these threats aren't slowing down. They keep increasing over time. And while in some ways, some places might say some of these technologies might seem like a luxury to the Alex, you know, because the pace of innovation that these threat actors are usually, you know, as Bernie said, you know, these kids can make ransomware with AI agents like crazy theory kind of true, right? It's actually really true. But, you know, we need to use the same tools at our disposal for our defense. And these tools are integrated into these Cisco networks and they come out of the box and able with these licenses. And the most amazing part about working in the Presidio is that I get to hear Bernie remind me that we're talking about focusing on enterprise networking here, Alex, which we've been focusing a lot on just campus switching. You know, Bernie's reminding us that those same licenses, those same switches, the same technology can be used in operating technology environments like factories in space of the take-roll products, other arms and machinery and things like that. It's a water supply and they are some of those risk oil and gas power generation. Absolutely. And you think about that, you know, Brittany, you bring up a great point. It's, so you think about years ago when we started talking about the conversion network. And so we technically are in a conversion network. We have voice, we have video, we have data. But what has stayed unconverged is exactly what you're talking about. The automation within our factory and our manufacturing because they're so afraid of, you know, well, if this happens, then that's going to shut down the process, whether it's a water purification, if it, you know, whatever the aspect, I work with a few nuclear agencies. And so, you know, having that get shut down, that's huge, you know, but being able to now have a surety because of features that are just going to be embedded within these switches can allow for a truly converged network across the board and ensuring that they're a secure, reliable network because I'll say some of those old OT networks aren't so reliable. They're delicate in the sense, but also crucial, right? You know, I mean, I have a vested interest in that area. My husband works in oil and gas and his job is inherently dangerous. And so without things like those sensors online and making sure that they are doing what they're supposed to be doing when they're supposed to be doing it, people like him could be at risk. So it's just too important. We have to, you know, we have to remember, you know, OT networks and all of this. And now this brings a whole new set of capabilities to the table and it's just very exciting. So, and I think, you know, you mentioned earlier, zero trust networking. I feel like at one point that was almost a nice to have or a buzzword, but now it's a half to. It is absolutely, you know, necessary. And so these embedded features are going to help clients on their journey to actually achieving that goal. It's no longer just such a far stretch. So, including and not limited to our OT and campus networks. So, that's right. Because that's the other thing. I mean, you're even zero trust our campus networks, you know, and being able to have somebody come in.
and plug in the device and you know, sure, you have, you have older knack style, you know, whether it's ice or whatever you, the flavor being. They're, they're a ways around it. And to have it now be implied in right at the access switch and be embedded within that switch to be able to put an SGT onto that device immediately and lock it down. Oh, well, no, you have this already inside your laptop. Sorry, you're not allowed on this network. You must, you must get allowed onto it and really just not trusting anything is going to be a game changer. And I, I think, and being able to secure and provide quality and reliable networking. So I couldn't agree more more. And when we look at all those things and all those features that you're talking about there, you know, it's really adding up to these more modern pieces of infrastructure to make your life so much easier, bring you up for more innovations. So I like, you know, no, all three of us network engineers, this call have been around here for a while. And I talked to all of our clients and I say, you know, how long you've been in your organization? What was the north like when you got there? And what is light today? And I frequently get to people who say, I've been here for five, 10, 15 years. And I said, okay, well, you know, how big was it when you got here? And they were like, oh, I got 20, 50, you know, 100 switches. And I said, well, how many you have now? And they'll say four or five times that much. And I said, so and you have a four or five times larger network departments. And the answer is never yes. Right? So you know, lower. Exactly. The truth is, Alec, the problem that we have in networking is we've been so successful as an industry, making bigger and faster and better networks. They're harder for us to maintain just because of how success we've been and how large they've grown. So the modernized networks are not just faster. They are faster, but they're also easier to use, easier to change, easier to operate and troubleshoots and easier to help enable our clients' businesses to help them succeed. And really, and that's one of the greatest things I can't out of Cisco live with because it's not just that these are new switches. It's really this new opportunity for innovation with our clients. I think Andy, honestly, I think you said it perfectly. I mean, there are so many reasons behind, you know, modernizing your network. I mean, we went into this conversation trying to kind of answer that question. Why modernize? And I think we answered it because it's not just that you're freeing up the time of your engineers. It's not just that you're doing, you know, doing the whole AI thing. I mean, AI is top of everyone's mind right now. So everyone's doing it, but it's not just because everyone's doing it. So many benefits there. So I think you wrapped that up. I think Andy, you like concluded it perfectly with that last point. So great job, guys. Thank you, everyone, for tuning into this episode of the digital decode. Please rate, review and subscribe to the digital decode wherever you're listening right now. Make sure to connect with us all on LinkedIn and follow the procedure and Cisco LinkedIn accounts to stay in the know on all things tech. Also, if you would like to expand your knowledge and dive into the Presidio Cloud journey, I highly recommend checking out Presidio Explores, which is linked in the show notes. Presidio Explores is our virtual adventure that will take you through a number of journeys that will have you scaling new heights in cloud collaboration, tune in next month for another episode. Navigating the digital transformation of your business is no joke. Presidio can help. Presidio is a leading IT solutions provider that helps clients simplify IT complexity and drive return on IT investment. By investing in the future of IT solutions, we stay at the forefront of technology trends, which ensures that our clients have access to a wide range of technology solutions. Let Presidio help you tap into the extraordinary potential of a digital future. Learn more at Presidio.com. You've been listening to the digital decode for more glimpses into the digital future. Subscribe to the show on your favorite podcast player. And if you like the show, give us a rating. Until next time. [BLANK_AUDIO]
Podcast Summary
Key Points:
Cisco announced new networking hardware (e.g., 9350, 9600, N9300 switches) at Cisco Live 2025, featuring embedded DPUs and enhanced AI capabilities.
AI integration in networking enables proactive operations, automated troubleshooting, predictive analytics, and improved security, moving beyond manual CLI-based management.
The distinction between "AI for networking" (using AI tools to enhance network engineering efficiency) and "networking for AI" (designing networks to support AI workloads like GPU clusters) is clarified, with emphasis on practical, accessible adoption.
Cisco Meraki dashboard and AI agents simplify deployment and management, making advanced features like automated patching, threat detection, and performance monitoring more approachable for clients.
Modern switches support scalability, micro-segmentation, and edge inferencing, addressing growing demands while enhancing cybersecurity through proactive threat mitigation and bandwidth management.
Summary:
This podcast episode discusses key announcements from Cisco Live 2025, focusing on new Cisco networking switches like the 9350s and 9600 series, which incorporate embedded DPUs and advanced AI functionalities. The conversation highlights a shift from legacy, manual network management using command-line interfaces to AI-driven, proactive operations. Experts explain the difference between "AI for networking"—using AI tools to improve network engineering efficiency—and "networking for AI"—designing infrastructure to support AI workloads, such as GPU communication.
The Cisco Meraki dashboard, integrated with AI agents, is emphasized as a user-friendly platform for deploying these technologies, enabling automated troubleshooting, predictive analytics, and enhanced security. Use cases include automated patching, threat isolation, and performance monitoring, helping organizations scale securely while mitigating risks like cyberattacks or bandwidth issues. The episode underscores the growing demand for modernized networks to meet evolving business needs, with Cisco's solutions offering both hardware innovation and software-driven intelligence.
FAQs
The podcast decodes and deconstructs digital technologies, strategies, and best practices impacting how companies deliver value to customers, avoiding buzzwords and focusing on practical insights.
Cisco announced several new switches, including the 9350s, 9600, new data center switches like the N9300s with embedded DPUs, and emphasized advancements in AI and network operations.
AI enables more efficient network operations through automation, predictive analytics, and AI-powered diagnostics, reducing manual tasks like command-line troubleshooting and allowing engineers to focus on higher-value outcomes.
'AI for networking' uses AI tools to enhance network engineering efficiency without needing extensive GPU resources, while 'networking for AI' involves designing networks to support AI workloads, potentially requiring specialized hardware like GPUs and higher bandwidth.
The new Cisco switches, such as the 9350s, can be integrated into the Cisco Meraki dashboard to utilize AI agents out of the box, making AI for networking more approachable and manageable through a centralized cloud platform.
AI enhances security by enabling proactive threat detection and mitigation, including identifying unknown attacks, isolating threats, and preventing network flooding or exploitation before they cause damage, moving from reactive to proactive defense.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.