 Good morning and bienvenidos a Barcelona. My name is Savannah Peterson. Join us here with Dave Vellante. We are on day four of our live coverage here on theCUBE at Mobile World Congress. We're excited, developers, developers, developers. We're bringing the heat, and we've got two fantastic guests joining us with a very interesting and exciting partnership. We're going to go real deep on the tech side. So for all you nerds, you're going to want to tune into this one. We've got Dave and Chooja. Thank you both for being here on the show. How are you guys feeling day four? I'm feeling great. First of all, we were here last year, right? On this very stage. Welcome back. And I think actually the mood's even improved further. It's like, I think last year, everybody was back from COVID and there was relief. This year, you actually feel palpable energy, enthusiasm for the industry, which is nice, right? And I think part of that's being driven by the AI everywhere. Like it's just, you can't escape it. I mean, there's robots attacking you in the aisles. I mean, it's literally, there's like all sorts of hovercraft and things like that going around. But it's a, but it's, I think people are seeing a future state, right? It's allowing them to sort of break away from, because clearly telcos have a heritage in infrastructure and being sort of hide bound in these big boxes and things like that and big iron, right? And now the AI is sort of giving them this idea of this horizontal, lightweight, loosely coupled future, much more intelligence and contextual. So I think that's generating a lot of excitement on the floor. I hadn't really thought of AI as spelt, but that was a nice, nice opportunity. What about for you, Shujah? Yeah, I mean, as Dave said about the AI is pretty much the hype this year. We have done a lot of work from our side as well in order to prepare since last year to this year to bring in AI's story across every single portfolio, technology, as well as work across the partner ecosystem in order to look at the use cases, the areas where you can jointly collaborate and deliver some value to the customers. So it has been pretty exciting so far. A lot of good partnership and collaboration and relationship across the whole broader partner ecosystem and some really good demos and demonstrations that could actually weave into the way of 5G monetization and other areas where telcos might be interested in. We want such a hot topic there with monetization. We'll dive into that in a second. Dave, can you tell us a little bit more about the partnership here? Oh yeah, absolutely. So we were here last year with Azar from Red Hat and we're talking about Telco SuperCloud. Yes. We love talking SuperCloud. We've got a whole new demo to talk about the presentation we're doing here at MWC24. But you know, this is actually a long running collaboration. We actually collaborated with Red Hat on the very first Etsy NFE proof of concept, right? In 2013, right? Back in 2013, we were already demonstrating a venture-driven model-based poly-control automation using graphs and agents and nobody knew what I was talking about. Yeah, so you know, things take time to, but now, again, AI's here and I think our story can be absorbed better. I mean, we've been succeeding through the industry but now I really can feel the wind behind our back because people want contextuality, because people want personalized experiences, because people want to optimize, well, they want to optimize their infrastructure too, right? Low energy, right? You know, low latency, right? This is another generation of technology, right? That's delivering this and Enterprise Web's going to be part of that. But the collaboration with Red Hat's very natural because you can just very simply state it is Enterprise Web Platform's the application layer. Red Hat's the infrastructure layer. Together, it's unified network and service management, right? I mean, we literally weave together so nicely, right? At all aspects across from OpenShift, Ansible, there's just so many points, I don't know if you want to bring in some of the others. Right, I mean, as Dave said, let's call it AI-powered super cloud with the joint collaboration. So Red Hat vision is build ones, deploy everywhere and operate autonomously and we have a vision of open hybrid cloud powered by AI. We have done a lot of work with our partner ecosystem horizontally and vertically in order all the way from Silicon at the OEM level to the networking stack, to the switching fabric, go up in the layer at the application layer with our workload providers and especially the one team that is very common here and is the abstraction layer because now we have this multiple operating model, multiple operating systems, multiple different cloud environments as well as different silo or mission that is going on in different areas of the business, like IT, network, cloud, I think that the foremost important piece is the abstraction layer, how do you abstract all of that underneath in order to expose that common and unified platform and that's where our partnership has evolved as Dave said, we're starting from the early days of NFE and our mission and orchestration into more meaningful and focused AI powered abstraction layer that could actually serve the purpose across, I mean the GSMA open gateway with the open APIs as well as the common layer of service management and orchestration that is being talk of the town in the open ran business or 5G core edge in the data center. So I think our partnership has grown up as well as our technology stack has actually grown up and plus we have some future vision in order to deliver some of that common theme across the next generation of the telco cloud. So let's get into that a little bit. You're talking graph and graph ops here at the show. Traditionally you think about graph databases, they've been used inside of places like security. They've not been widely applicable in the enterprise because while they're expressive they're really hard to query, you don't get that simplicity of query that you get with SQL. So you actually have to go back 10 or 15 years to query graph databases. That's beginning to change. It sounds like you've got, I really want to hear more about your applying, I mentioned security, it sounds like the network can really benefit from graph and graph ops. Can you explain that? And then I actually Dave want to get into what that means for the enterprise outside of telco but we're here so let's talk about it. We're right back to abstraction, this is a subject we talk about all the time. And I think if you really think about it in software abstraction equals power. When you build silos, you build use cases. You build static things. And every time you build a static silo thing it's the next problem you have for interoperability and end automation and global visibility. When you have abstractions you're essentially going over the top. And that's what we originally talked about super cloud. So you're spot on regarding graph. The history, graph goes back to Aristotle in objects and things. A little bit of history. A little bit of history, yeah. I thought I was an industry historian, wow. In mathematics it goes back to 1736 bridges of Koningsburg theory, seriously, I had no joke. Love this knowledge right now. But in the last 10 years or so the graphs that most people know about the semantic web, label property graphs, RDF kind of things like that triples those are really data graphs for data scientists and data analytics. They are not operational databases. And I think people are looking to graphs because they see graphs and they see flexibility. They see things that you can analyze very quickly and they want to bring that in but there's an impedance mismatch. A graph database as practiced by semantic web they're great for what they do. That's their use case. But they're not going to be, so enterprise web is using something called hyper graph. So we actually have, we just got an announcement of our 21st awarded US patent. So we have 21 awarded patents on the use of- Congratulations. Thank you. It's exciting. Yeah, it is exciting and expensive by the way. Yeah. That's what it is. You know, it's a couple million I guess spent on other things. More flames. Yeah. But, you know, hyper graphs essentially are an abstraction of graph itself. Essentially enterprise web instead of actually building a graph database which actually as the complexity of your objects and the complexity of your domain grows you have to, you were saying sharded over a lot of databases and then you have this, your reads and your writes become much more expensive and you have these problems called cascading update problems. I mean they're real fundamental problems. They're roadblocks to using it for anything real time. Enterprise web essentially is almost like old fashioned list processing. If you're a Lisp, like from the 1950s, McCarthy, right? Where essentially enterprise web is just like a long skinny table and we actually project the graphs instead of having a database that models a graph physically we actually just run as rows. And what we have is tags on the rows and those tags are relationships. And what we could do is we can read those tags and project graphs and say, okay in this context these things are related. In another context these things are related. And what it means is that we're always interpreting the graph based on the moment, the context of the moment. Enterprise web is driven by that. There's nothing else enterprise web does. It says there's a request, a query or command for something. It's coming from Dave. It's coming in from Barcelona. It's coming with what's called your in-band metadata. And we take that little bit of in-band metadata and then we throw that against it. We use that to sort of bootstrap a walk across the rest of our graph. And then we say, we know all these other things about Dave Vlante. Let's bring those into bearer and we'll apply security. We'll apply all these policies and we'll bring it down and say, here in 200 milliseconds, Dave is a completely personalized response to you. If it was for Shujar in the response and he asked the same question, but in his context, in his role his response might be completely different. So you infer that metadata, infer from that metadata and now you can apply AI to do a lot of that heavy lifting. And actually we can work very like sympathetically with AI too, because graphs naturally work with AI because if you think it, if you have a SQL siloed solution and a hierarchical rigid database, it's hard for the AI to really roam that, right? But graphs are based on relationships and because we're a hyper graph or multi-dimensional, we can expose that graph through an API to AI or to LLMs, right? And they can essentially walk our graphs because any form of AI is looking for patterns. They're just different methods to detect patterns. Well, relationships help you identify patterns. So we're sort of on the deterministic logic-based side driving those real-time operations and then we could work with AI to get those higher level inferences. So a good example is last year after MWC 23, we met with Microsoft and when we, a few months later in May of last year, we demonstrated the world's first telco-grade generative AI for intent-based orchestration. Nobody else- Casual. Yeah, so we did that directly with Microsoft Telecom. We were the first people in the world to actually show that telco-grade orchestration within LLM, right? And it's nine months, it's almost a year later and nobody else has matched it yet, right? It's partially because of our design choices, it's our architecture. Enterprise Web is really sort of suited for this moment, for this AI-enabled, intelligent, real-time, you know, 10 years ago, people thought I was crazy, but now I think, you know, like, you know, who's going to want that? Do you feel vindicated? Do you feel validated now? I am, Dave, I don't think 10 years ago, people thought you were crazy, other than the fact that they thought what you were trying to do was impossible, so maybe they thought you were crazy for trying to solve that problem, but it's a real problem, right? And you think about, well, you think about packaged apps like SAP or Oracle, they define how you have to operate, and they essentially, you call them, I think, use cases, and that use case is, all the data is locked inside of that use case, so what are the advancements in software technology, this is what you're at the forefront of, that are allowing you, enabling you to span, to abstract across those environments so that you can now have a more logical layer of your business, a representation of your business, can you describe that technology? How much time do we have? But, so, you know, in short, so Enterprise Web, it's really, you have to almost back it up for a second, you could, 10 years ago, you could see distributed systems coming, right? Even before the cloud, right? Cloud's not to 2012, 13, right? And that's still nascent, right? Kubernetes is not to 2016, we think it's been around forever, but 2016, that's not that long ago, right? But you could see as the industry was fragmenting, right? And systems were getting, people and systems and information were getting distributed. In a distributed system, it's just logical, and this was my original sort of insight that when I founded the motivation to found the company, was to say, you know, actually, in a distributed system, you don't want to build silos, silos are going to be just pain points for interoperability and change, and we're just going to be exacerbating all our problems when we go to the cloud, if we keep our silos, you want a horizontal abstraction over the top. You want, even though the world is increasingly dynamic, distributed and diverse, you'd like one layer where you could say, hey, all these objects look the same to me. And in this place, I know they're unique, I know they're all snowflakes, but in this place, they've been modeled, mapped to a DSL, they've been mapped to this graph, and I could look at them in the same way, I can discover them the same way, I can compose them declaratively, and I can deploy them and manage them with central policy management. That's really critical. It's a better user experience. Yeah, and what's really happening is that's all abstraction, because in the reality is enterprise web's runtime, anytime you abstract something, what you're doing is, when we were talking about developers in the beginning, right? Yeah, yeah. So, developers, so for the developers, right, developers want to solve business cases, they want to solve business problems, and the way to do that is remove barriers from them. If we can create an abstraction, say, look guys, everything's in a catalog, forget the fact that they might be unique, and they have idiosyncrasies and all these details here. In this catalog, you can find anything you need, a router, a switch, whatever else it is, right? I want to go there, I want to compose it into the service that I need to compose for my job. I want to set the SLAs on it, the policies, and then I want to deploy it. I don't want to care about how it runs. I don't want to care about where it runs. I don't want to be able to tell the system. I want it to work. I want it to work. I want it to be low latency. I want it to be energy efficient. And I want the system, so every time you do an abstraction, what you're doing is you're raising the abstraction so the developer and his tasks are simplified or his, her, right? So their tasks are simplified and automated, right? And what you're doing is when you're transferring the burden onto the runtime, the runtime that takes responsibility. How we do that, we use agents. So Enterprise Web is so the graph is where you get to the model and the execution, the runtime of Enterprise Web is completely agent-based. Enterprise Web is a serverless runtime. It's part of our patents. It's not Amazon Lambda. It's our own implementation of serverless, which just means that we have these agents that live inside our graph as well and every time something was we were talking about, there's a queer command. We just fan out agents to customize that response for you. So imagine if you went to the car wash and 20 guys came out to hand wash your car, right? They just come out of the building and they just swarm your car and they just do, and then they're gone. And then they're on the next car, right? What we have is a swarm of agents that come out. They do the tasks they need to do and then they go away. They're completely stateless. There's no long running threads. There's no long running sessions. So if you look about- You don't have to manage them, all right? You don't have to manage them as an IT person. Yeah, yeah, yeah. Because enterprise web is doing it. Yeah, if I may add just to build on top, because everything has to run on the infrastructure, right? In the past, 10 years ago, was not a common vision across people, processes, and a platform. There's some harmonization, but now I would say there is more broader understanding of this common and unified platform across every single environment because today you could see that there's a lot of edge development that are happening, edge private 5G, that you have this core network, then you have the data center. Now this open run has come in. So there are different set of technologies, different environments across any cloud, public cloud, multi-cloud. You have this small application cloud. All these environments are actually getting mixed. So now you need that form of uniqueness across your infrastructure layer. You need that same uniqueness across your abstraction layer. Your graph ops, as you explained, you want this single pane of glass that could actually provide you the capability and ability across any single environment. And that's where Red Hat brings in the technologies that actually goes all the way from device edge in a small box to the core and the data center. Uplift enables some of those technologies which we have integrated successfully with enterprise web graph ops in order to bring in this capability of OpenShift AI for example, some of the AI tools in order to train the model, in order to deploy the model, in order to continue this reinforcement. So all those capabilities that has actually been added in the platform in order to uplift and enhance the user experience on top. This is where the partnerships really starting to come into focus because essentially you're an integration layer across legacy applications. You're taking what were previously islands of automation and you're spanning those. Now I can automate my entire business. So I can now, we're seeing the days where you can get a digital representation of your business in real time. People, George and we talk about this all the time, people, places and things, Uber for the enterprise. Imagine an Uber like experience except it's your enterprise. Everything that's happening and you got your agents swarming and then leaving and then you've got the graph representing that business. Yes, yeah. So it's maintaining the state of the business too, right? Which is another thing that graph databases wouldn't do naturally, right? Because when you're dealing with that kind of complexity you're literally managing the state of every user interaction, every system interaction, every deployed function. You're managing a lot of, so the run time takes a lot of responsibility to manage the consistency across all the partners. So everybody looking at an object at any given time, state has to be consistent, right? Absolutely. But that should be lifted up from the developer, right? It's just too much, right? The fact is the demand for integration and interoperability and new applications is going through the roof, right? Humans don't scale that well, right? We need to make it easier for developers to do these tasks, right? So they can accelerate service delivery, right? Deliver new products to market faster, connect their end to end solutions and evolve them over time, right? It's another thing about graph, just like your social graph on LinkedIn, right? We're connected on LinkedIn, right? So on LinkedIn, you'd like to think you're only adding people but you can add people, leave people, right? And just like, you know, you have friends of friends, friends of friends of friends, you have this, there's a graph behind you, right? And that's evolving with you over time, right? And that's interesting. So Enterprise Web allows you to do that as a business because that was my original motivation. I am, at my core, a business guy that likes to get technical. I'm not a technical guy. So I want- You haven't noticed. I wanted to solve a business problem. You're not a technical guy? No, I'm talking about- You don't build my technical guy, my silent partner. But the, because I wanted, I saw it as a business problem. It's like, I actually got mad at that same thing you were talking about. So back in 2010 or so, I was thinking like, hey, I'm a business guy. I want to look down my people information capabilities. I want to flexibly compose them for purpose. And I don't want IT to tell me I can't. So I, nobody was doing it. So I figured, okay, I'm going to solve this problem. And I did. So far, so good. You said something while we were getting prepped and we could go on all day. So we're going to have to wrap this up here in just a second. However- Why don't you just have them bring drinks here? Yeah. I mean, I'm here for that. A little early. There's, you know, whatever. It's day four. There are no rules here. We're in Barcelona. You mentioned, I'm a big fan of The Edge. We've had quite a few great conversations on theCUBE this week about The Edge. You said that you can do things nobody else can at The Edge. Explain. Oh, you're going to call me out on that? Well, now I want to know. Now I want to know what those things are. You teased me with it before we went live. Okay, so then it goes down to like the characteristics of our platform itself, right? So we already agreed that people want to be real-time, intelligent, contextual AI. So we agree. So now imagine if I told you I could do that all in 50 megabytes and it could run at the core, at The Edge on a device. Ooh, yeah. Our full footprint is 50 megabytes. I can actually cut it. Our kernel is only like six or seven. Really? And when I talk about 50 megabytes, that's an entire telco model. So the ran, the core, the transport, layer one through seven, I can reason over that entire space space that's something nobody else can do. Yeah. And see, this is a big thing that we were talking about, trapped use cases and stuff like that. AI is all about reasoning, right? It's making intelligent decisions. Like if you were a doctor, you would look at me, if I was on the operating table through the emergency room, you would look at me holistically. You would just say, Dave's got a scratch over here. Let me just treat that and you'd miss the gunshot wound. Right. Or you'd neglect to find out that I have allergies to some drugs or some issues. You would want to know my chart. You would actually be managing me. You'd be checking my blood pressure. You'd be checking. You'd be taking all these facts in all the time to make sure the actions you take are optimized, right? If you do everything siloed, you've just contained what your decision space is, right? If I'm managing a service or a slice and I'm all the way up here, it's a network service, I want to understand what's the RAN doing? What's the core doing? What's the transport doing? What are my functions doing? Do I need to dynamic? What do I need to like configure here and what do you need to configure there to make sure this thing maintains its SLA continuously, right? That's a sort of next level problem. The fact that we could take that and we can bring that down into a sub 50 megabyte footprint. Yeah. Right? Yeah. That's game changing. Impressive. Yeah, thank you. Okay, so you weren't lying to me before we started the show. That's awesome. Shuja, Dave, thank you so much for being here on the show. Absolutely. Oh, it was a pleasure. It was a discussion. Dave, thank you for always keeping it entertaining and both of us getting technical on graphs here. I love it. And thank you all for tuning in to our four days of live coverage here for Mobile World Congress in Barcelona, Spain. My name is Savannah Peterson. You're watching theCUBE, the leading source for enterprise tech news.