RSA Conference 2023 | Dave Duggal | EnterpriseWeb

theCUBE.net

More from RSA Conference 2023

Dave Duggal

Founder & CEO

EnterpriseWeb

Info

Dave Duggal | EnterpriseWeb

Dave Duggal

Founder & CEO EnterpriseWeb

search

>> Welcome back to Moscone Center, everybody. This is day four of theCUBE's continuous coverage of RSA 2023. I'm here with John Furrier. Dave Duggal is in the house, again. He's the founder and managing director of Enterprise Web Dave, we just saw you at MWC in Barcelona. That was an awesome show, I mean, bigger than this, but this is big.

>> Yeah, this is big. It was about 90,000, 95,000. So it's good to see these conferences coming back. And MWC feels like yesterday, right? I mean, seriously. I mean, I know you guys are on the move constantly, but, but we had a great conversation with, at that point, you know, John was, you weren't there, but was it was Dave and yeah, Telco Supercloud with Dave and Lisa, right? That was fun. We, we, I think we put on a real good show there, but the, you know, there we were talking about how the, you know, telcos want to work across their domains, work across layers, right? So they can flexibly optimize services for low latency and energy efficiency, right? The only problem is that they're static, vertically integrated, tightly coupled stacks get in the way of their digital business transformation, right? So they need to be much more, you know, dynamic, loosely coupled, horizontally architected. And that's what we do at Enterprise Web, right? We have a intelligent interoperability and automation platform that allows things to be just highly dynamic to, to meet these kind of modern needs.

>> Yeah, and the thing about our, our conversations all week, Dave and I have been talking about this, is the platformization of security, which is kind of their first awakening to like, hey, let's build a platform. But the difference is, is that what you're doing is essentially like a real platform, right? They're, they're doing focused insecurity. But the conversation is, if you don't put network and security to enable cloud native then it's, it's not the right fit. So what we see security doing is saying, okay, if security and network run things, it then has to enable on top, this is kind of what your telco, cloud vision, Supercloud vision is.

>> Yeah.

>> Talk about where your platform intersects with some of the narrative coming out of this show, and what does that mean?

>> Yeah, so I think the whole, and you're right, platformization is a thing now. And I think that's actually almost a maturity of, you know, the move to the cloud, right? Because now people are thinking of distributed systems as systems. This is systems thinking. That's what platform is, right? It's a holistic view of your problem space. Your problem domain, right? You know, people used to do these siloed, specialized products. And the problem is, the problem is no longer siloed.

>> Yeah.

>> The problems, you're a whole attack surface, right? So you, you got to go wide. You have to have an understanding of that domain. And, you know, and one of the other hot topics here, of course, is a generative AI, right? We came back from Barcelona and all the sudden, generative AI. It started stolen the thread, right?

>> We've been joking that it was invented over the holidays. AI.

>> Yeah, absolutely. It is sort of funny cause it's been kicking around for a couple years, but boom, it just exploded, right? I mean, it's sucked all the air out of the room. And, you know, the funny thing, it's, it's great for us, cause, you know, Enterprise Web, given our nature, is the ultimate serverless backend for generative AI, right? And we're already working with the big players here because if you think of it like, this is a revolution in data analytics. That's what it really is about being real time conversational, interactive. Well then you want your backend to be able to support that. If your backend's sort of static, you're not going to optimize against this information flow that's constantly coming in, right? You want your processes to be able to react to what's going on in the real world. You want your transactions to be optimized for that as well. Your operations to be synchronized. So, so we're, we actually believe not, we already have a, an industrial grade, no-code automation demo using generative AI. We just started, I've been here all week in the valley running up and down the, the usual streets, right?

>> Yeah.

>> And it's blowing people away.

>> Dave Vellante: So what are the requirements, Dave, from the backend to support the sort of new wave of generative AI workloads? What, what are the, what are the fundamental characteristics of that backend? What do they have to look like?

>> All right, so, I mean, obviously this goes pretty low, you know, but, so I think starting at the highest level and then working down, because I, I don't want to lose people too quickly.

>> Oh! Our audience,

>> I've been accused, I've been accused of that before. And but at the highest level, you can say this is two sides of the same coin, right? This is just the latest in advance analytics, meaning the latest in advance in automation, right? That's, that's really what it is. And those two have always gone together symbiotically, right? Data drives process, process creates data drive, right? And it's just a loop, right? And that's what you really, in a pure sense, that's what you want. You want to be able to observe the environment, use those observations to drive next actions. You want those next actions to be really fully optimized for what you, you know, based. But you also want them to be safe and secure. And some of those are some of the issues with generative AI right now, right? You want them, you know, setting aside the whole IP issue about what generative AI's trained on, that's one issue totally sort of orthogonal. But the other issue is generative AI is not always right. Sometimes it's adamantly wrong, sometimes it hallucinates. I mean it's, you know, it's working over those large language models. It's essentially drunk on data, right? So you need to put constraints on top. And that's the difference. Again, that's the historical difference. That old Olap versus OLTP, if you remember those days, right? Analytics versus transactions, right? Probabilistic statistical on one side, deterministic rules based on the other side. You could think of those as almost being two sides of the brain, right? Left brain, right brain. You got sort of this creative side that's constantly reacting and then, then you have sort of the logical side of the brain that's actually has models and it's applying rules to things. Enterprise Web is bringing sort of that industrial grade side to it. If you look at a lot of the automation demos for generative AI, you, they're clearly experiments really like rough beta prototypes. They even say it when they're doing it. And they're sort of like, if this and that kind of stuff really stuff. And they're not thinking about security and they're not thinking about IT governance. Well guess what? Nobody at the enterprise is ever going to touch that, right? You know, the enterprise needs, I mean, this is, you can't have creativity in your transactions, right? You can't, right? It's like, yeah.

>> You don't want yourself,

>> It's got to be binary.

>> Yeah, you like your banking ledger to be, you know, hey, you know, it's like put in a hundred, it shows the a hundred, right? You, you, you want that kind of transactional control. And so Enterprise Web is adding that, let's say rigor, right? So, so in this demo that we have today already, we can demonstrate it working, it's great. We're essentially, it's a Microsoft stack we're working through today for the obvious reasons, it's really the one that's breaking away. But we're working through Jarvis, which is our NLP, right? So we can actually talk to Enterprise Web Jarvis through open AI, right? To enterprise web. There's another product in between I'll get into when I go into the technical details that you ask. But essentially I can talk to Enterprise Web and say, I want a network service, something really advanced, right? That's even hard for an engineer. I want this, please compose this for me. Then when it composes it, I want it to be optimized this way, configured this way, and then deploy it on the network. And this, in a couple of minutes, I could dictate this to the machine. The machine does it and then it shows me the state, right? So I can compose, deploy, and manage on an ongoing basis the day zero one and two working with generative AI.

>> Dave Vellante: So you can say, okay, optimize for low latency and give me an estimate as to what it's going to cost in a monthly.

>> Dave Duggal: Yes.

>> Okay, that's a little too expensive. So dial down, dial up the latency a little bit, and you can speak to it.

>> Absolutely.

>> That's your language way. And then optimize your infrastructure for an outcome.

>> Yeah. And energy, and this is exactly, you know, in telecom they're calling that intent, right? [Dave Vellante] Yes.

>> Managing the declarative intent, right? So a developer can say, you know, energy efficiency, maybe it's not latency, maybe it's energy efficiency, right? That's my, so manage, I don't want to care how you manage energy efficiency. The details of that are way too low level. You know, it's down to the network, it's down to the, the RAM, the cores that you're touching a lot of things at that point. So, but you know, does a business really care? No, it just wants its SLAs and SLOs met, it wants its intent met. Honestly, if you think about we've spent the last 10, 20, 30 years of computing thinking so much about our tools, right? Completely independent of the business behavior and objectives and now we're actually almost at that stage, right? That inflection point where we're really going to enter the 21st century for the first time, right? I'd argue that we haven't been in it, right? We're still doing 20th century automation now. We're also going to be in, the focus will be you talk to the machine, you express your intent and the system in onset and it's fully traceable, right? It's fully governed.

>> Talk about the self-healing aspect of, I noticed that was in your LinkedIn post.

>> Yeah, yeah.

>> Because that's the, that third wave of gen ai. You have prompt engineering, prompt ops, prompt tuning, which is, let's see the query operationalizing it.

>> Yeah.

>> And then self tuning, tuning it, which means, yeah, it's on its own. Where do you guys fit in on that piece? Cause I can imagine that's where, when you give an order into the command provision, this, the hallucination side's a concern. So I'm assuming you're doing that self tuning, self-healing. Take me through that piece of it.

>> So, so we do this, so, you know, we're using the Microsoft Gen AI in front, right? So that's the, that's, let's call that the, the Jarvis and the Open AI up front. We put another technology between us. Essentially what you want is a vector-based time series database. So this goes to, to your question actually the both of your questions, because we do this cause we want an intermediary, we want to be arm's length, we want a nice separation of concerns between us and the generative AI for exactly the reason you just described, right? I'd want that hallucination to flow into my commands, right? So what what happens is the hallucination is working through this time series database, which is putting that data in order, right? It's taking this massive stream of event information and we're, it's applying rules to it and then just throwing events to enterprise web, right? But those events are already sort of filtered in a way that Enterprise Web can respond to them. And what it's doing is that, and this we work with a company called KX, which a really interesting company, you guys should have 'em on some time actually.

>> What's the company?

>> KX

>> KX, kx.com, 20 of the top 20 realtime.

>> They're AI ML partner, right?

>> Yeah, yeah, exactly. Yeah, so 20 of the top 20 financial institutions use 'em for realtime training, right? They wanted to enter telco. So they actually partnered with us because we're, we have tech telco domain expertise, right? In this case, so now totally different use cases and more, cause General Enterprise Web is generalized, generative AI is generalized, KX is generalized for analytics and everything like that. So we use 'em as an intermediate. So what happens is they're, we're arms length from generative AI when it's talking to KX, KX is then translating it into the symbolic language is semantics that we speak for us, right? And it's almost like a universal translator, right? Translating back and forth between Vector that.

>> Brings coherence.

>> Coherence, right? Because essentially what we're allowing it to do is the, through the intermediary generative AI is walking our graphs, it's walking our domain models and introspecting our catalogs. So when I actually, when it, when I call, talk to my system and I say I want something, it's literally walking the enterprise web graph saying, oh, I found it. Oh, and here's the object I want to use. It's a Cisco router, a Juniper router, or whatever it is. I'll compose this with this and this and this, and maybe it's a fort net security firewall, you know, whatever, whatever it is. It'll put that together, apply my constraints and then implement it. The controls come on the Enterprise website. Completely discreet. So generative AI is enabling the conversation that's super powerful. Almost think of it as the most advanced user interface you can imagine, right? It's now Star Trek, right? This is, you know, all right.

>> We said in theCUBE one time, everything in Star Trek will be invented except for transporter room and that automatic food making.

>> Yeah.

>> But you just gave an example of whatever, a Cisco router or, or Palo Alto firewall, whatever. What is it that allows you to not care? And then is that different in telco?

>> So it, it, it's interesting. So part of it's what we've already had the sort of core enterprise web platform in enterprise web, essentially we have what's called an upper ontology, right? Upper ontology just means it's the, the generic set of concepts and types that runs across all enterprise businesses, verticals, whatever. Every business has people, it has organizational units, it has facilities, right? Those kind of things. And then, then of course, from a cloud and distributed systems perspective perspective, there are common types, right? There's formats, there's protocols at the upper level, protocol enterprise Web maintains that. It's sort of the universal concepts that apply to everybody. Then Enterprise Web allows you to model a domain, like whether it's telco, whether it's life sciences, whether it's IOT, it doesn't really matter to Enterprise Web. We enabled you to rapidly model a, a complex domain. And then what happens is you're onboarding objects into that domain, right? So that router you were talking about, that firewall you were talking about, right? So what you're doing is, instead of doing that old kind of point to point integration you used to do in the stack, what you're doing is actually just mapping up to those graphs. You're just mapping in metadata. So we're saying, hey, we understand Juniper Router or Cisco Router or Palo Alto Security firewall or Fort Net. We understand them completely in metadata. Enterprise web is a hundred percent configured by metadata, right? All enterprise web is a graph model, which has policies, which has metadata policies, you know, and types. And then it's stateless functions. And Enterprise Web uses that metadata to efficiently hydrate context for stateless functions. That alone is a super big idea cause serverless.

>> You have a very big idea. So the super cloud aspect of the telco thing, is super impressive. What you're getting at now is the modern platform vision. It's, enterprise is hard. We had Jeremy Burton on, who was a very distinguished executive over the years, now he's doing an observability startup. He's like, it's hard to do a startup, especially in the enterprise, takes years. Where are you guys at?

>> Tell me about it.

>> Take us through your progression because you guys have a solution now. How does someone get in? How do they get involved? What's the deployment? How do they get, how do they adopt the technology?

>> Yep.

>> Cause it is definitely in line with what people want.

>> Yep.

>> Some people don't have the skills. How do you guys do the managed services?

>> How does someone engage with you guys?

>> Yep.

>> And, and and how long have you been around?

>> Yeah. Fantastic. So, so I've been,

I've been personally? >> No, I've been around for a long time.

>> No, Enterprise Web.

>> So actually, so I had, I had started turnaround growing several companies in a row. And, and so around the time of 2008, 2009, you know, pre, pre before cloud took off, cloud doesn't really take off till 2012, before Kubernetes, before microservices, right, before containers. My previous engagements, my roles were, I was like seeing that the 20th century automation tools are getting the way, the kind of things I wanted to do already, just from a business intuitive perspective is like, why can't I do, why does it tell me what I can't do? I'm the business guy, I want this done. Why is being, you know, when I asked some behavior to be implemented, I don't say do that in concrete. Right? I'd like it to still be adaptable if my, you know, if all of a sudden my needs change the, the competitive landscape. So I start, I, in 2007,8, 7, 8 I read three, 400 academic papers and industry papers. I seriously, I actually, I left my prior job and I said, you know, I'm going to take off. I'm going to do sabbatical, I'm going to figure this out. And so I went out and just did all this research and just look it up, like what, what, why, why can't people be more dynamic? And basically it was the old, this old software was based on the old constraints of hardware and infrastructure. And when the hardware and infrastructure kept on advancing the software didn't.

>> Yeah, that's.

>> Right. You know, partially because those models of selling packaged software were pretty profitable. Selling lots of different components.

>> Highly motivated to keep it the way it is.

>> Yeah, they didn't want to capitalize it, right? So I said, you know, actually this should be horizontal. This is really capabilities architecture, right? I should be able to say, I've got a catalog with my capabilities and I have business logic over the top. My middleware should be this thin horizontally architected layer between my business logic and my things.

>> Yeah. >> And that's all that matters. I don't want to care how it deploys like we talked about earlier. So I started doing that and essentially I just work behind the scenes. I just self-funded it. I, I've taken no money. I've bootstrapped this whole company this whole way to profitability, right?

>> Congratulations, that's awesome.

>> Thank you. Awesome, yeah. Not an easy road, I don't recommend it to everybody.

>> It is, well it is, is an elite, it is an elite title to have that self-funded start.

>> Now you have options like you that, that you wouldn't have had if you done,

you'd taken money >> Well done.

early on right . >> Thank you. So we're in production, some of the world's largest companies and we go to market new channels, so.

>> What kind of company, how big are they?

>> Well, Red Hat, so last, on our last show we had Red Hat, you know, Azhar Sayeed from Red Hat was on with us. And that was a great show, so Red Hat looks at us as, you know, maybe being the application layer over OpenShift right. We have partners with Intel. We just did that award-winning test bed, five grand. We're, we're the closing loop with our partner AKX on the AI ML side, closing the loop from reported event to mediation.

>> Yeah.

>> In 300 to 400 microseconds, not milliseconds. The state of the art orchestration is 11 milliseconds. Alright. I'm at three, right. I'm orders of magnitude more performant they are. And the reason is I have no cruft, right. There's nothing but a graph and the stateless functions once

something happens. >> Yeah. We flatten that graph.

>> You got to love the land, >> Execute the functions.

>> Got to love them. Got to love serverless.

>> Yeah, oh, oh, and we implemented it. So we have our own, so we have 19 water patents, multiple patents pending, right so we actually, cause we existed before, there's other technologies because you have to remember.

>> At least she's throwing money at you right now.

>> We're getting a lot of interesting conversations. And the generative AI thing is, without me saying anything, the biggest players, and you know, there's only two or three, the biggest players in generative AI are super interested because right now every organization in the world is reconsidering their operations in light of generative AI.

>> Totally. Right. In the last two months, you.

>> If you're not, you better be.

>> And their concern is about the hallucinations which you're trying to address. That's the opposite of like, I don't, I want pure, reliable, scalable.

>> Deterministic.

>> And it's got to be portable.

>> Low latency, high performance, portable. Right, we're 50 megabytes, you know, again, we got all these great characteristics because we rethought it, right. And people didn't think it was possible so they didn't try. And also I think one benefit.

>> Well I think the architecture you laid out also business apps essentially could be put cloud right there. So you actually had, the architecture was perfectly lined with how the cloud spun.

>> Yeah.

>> Hence our super cloud and our rendezvous. Our serendipity meeting you.

>> Yeah, yeah.

Exactly. >> That's awesome.

>> Yeah, well actually, even before we went on on theCUBE with you, I went on theCUBE with you. You, you were already talking about Super Cloud Paws. So I saw that, I was like, ah, I got to follow up with

these guys. >> Yeah, yeah. We're in the same tribe, let's go.

>> These are my peeps. >> Yeah. Yeah, well I'd been on theCUBE back in 2015 in a structure conference with George Gilbert.

>> I remember that.

>> Back in the day.

>> I do. You still working with George?

>> Yeah, actually I've been having some conversations with him.

>> He's a great guy >> And he of course dodged me.

>> I remember Digital, digital Ocean back then. But they, I interviewed those guys too.

>> Okay, yeah.

>> Yeah, it was good time.

>> Yeah, so, so we go through market, through channel. So we have SI people like Tech Mahindras, one of our SI's you know, Red Hat, Intel, you know, Amazon.

>> How does someone engage with your customer? What, who wants to engage with Enterprise Web? What do they do?

>> Well so I mean, anybody can,

>> So a channel partner or,

>> Yeah, so they can inter, they can go directly to any of those channel partners or they can obviously David enterprise web dot com,

can I do that? >> Yeah.

>> Am I allowed to do that?

>> Of course, yes.

>> David enterprise web dot com, so they, anybody can reach out to me and, and I'll direct them to the right partners that can do service delivery and sport. I'm focused on building the world's greatest software company because I think the world needs it. Every epoch, right.

Every period >> Yeah.

>> Requires that new IT company to come and do that, and, you know, maybe it's Hubris, right? I'm a guy who sits in Glens Falls, New York and mostly working my t-shirt and jeans. But you know what, actually every once in a while the world, the ocean does need to be boiled cause it's rancid, right. And people need to rethink things.

>> Bill Gates start was in Seattle. No one thought anything about Seattle except,

same with Jeff Bezos. >> Yeah.

>> We're going to put Glens Falls New York on the map.

>> I love it.

>> And so, yeah. So it, it's been a journey, you know, it's been a lot of fun, I mean, I've learned a ton, right? I've become a student of organizations, you know, working with all these big players. I, I've been squashed so many times I've, I've heard every naysayer, but you know what? I knew our ideas were right.

>> Stayed the course.

>> Yeah, and, and that, you know, you know, luckily we had enough friends, partners, customers to, you know, support us. We knew that we were doing things right. We were validating it in the background. And that's what I knew I had to do. When you make big claims, you need big validations.

>> Yeah. >> So that, that's why Well, you got customers, you got this,

they're in production. >> Yeah, yeah. Big customers.

>> And your next steps now is what? Scaling to go to market,

>> Scale out. Scale out, this is ready to rip, right. This, this is, this has been proven and,

>> And you need funding to do that.

>> Yes. So yeah, we're at that scale. So we're at that inflection point. And so,

>> So growth capital, you're ready to go.

>> Yeah and we have other conversations too, but yes.

>> And you have, yeah or right, okay, but, and you have options because you,

you own the whole thing. >> Yep.

>> That's awesome, congratulations.

>> Thank you very much.

>> The gen AI piece is beautiful, so congratulations on that.

>> The, the gen AI almost, you know what the thing is partially, I think is now this is really our moment. I think we were on the right track anyway, but middleware's just never been sexy, right? It's been, it's plumbing. So selling middleware is a thankless job, right? And, and if we do our job well, it's like a utility, you shouldn't even know it's there. Gen AI makes us super sexy, you know,

being the back end to Gen AI >> Yeah. is a, we're talking about what, a hundred billion trillion market opportunity.

>> Yeah. Well we talk about it all the time that Amazon turned the data center into, into an API. And generative AI is going to turn the way in which we interact with technology into language

>> And it's automation compatible. People know what automation is on DevOps and operations. They see value in automating undifferentiated tasks. So I think it's, it'll, I think it'll play well. And then then having this more self-service, provisioning,

self-healing. >> Yeah.

>> All right guys, we got to go. We're, Dave always pleasure to see you.

>> Fantastic.

>> Great seeing you come out.

>> Great seeing you, keep right there. We'll be back with our next guest from RSA 2023. You're watching theCUBE from Moscone West in San Francisco. Right back.

>> Okay.

Dave Duggal | EnterpriseWeb

search

>> Welcome back to Moscone Center, everybody. This is day four of theCUBE's continuous coverage of RSA 2023. I'm here with John Furrier. Dave Duggal is in the house, again. He's the founder and managing director of Enterprise Web Dave, we just saw you at MWC in Barcelona. That was an awesome show, I mean, bigger than this, but this is big.

>> Yeah, this is big. It was about 90,000, 95,000. So it's good to see these conferences coming back. And MWC feels like yesterday, right? I mean, seriously. I mean, I know you guys are on the move constantly, but, but we had a great conversation with, at that point, you know, John was, you weren't there, but was it was Dave and yeah, Telco Supercloud with Dave and Lisa, right? That was fun. We, we, I think we put on a real good show there, but the, you know, there we were talking about how the, you know, telcos want to work across their domains, work across layers, right? So they can flexibly optimize services for low latency and energy efficiency, right? The only problem is that they're static, vertically integrated, tightly coupled stacks get in the way of their digital business transformation, right? So they need to be much more, you know, dynamic, loosely coupled, horizontally architected. And that's what we do at Enterprise Web, right? We have a intelligent interoperability and automation platform that allows things to be just highly dynamic to, to meet these kind of modern needs.

>> Yeah, and the thing about our, our conversations all week, Dave and I have been talking about this, is the platformization of security, which is kind of their first awakening to like, hey, let's build a platform. But the difference is, is that what you're doing is essentially like a real platform, right? They're, they're doing focused insecurity. But the conversation is, if you don't put network and security to enable cloud native then it's, it's not the right fit. So what we see security doing is saying, okay, if security and network run things, it then has to enable on top, this is kind of what your telco, cloud vision, Supercloud vision is.

>> Yeah.

>> Talk about where your platform intersects with some of the narrative coming out of this show, and what does that mean?

>> Yeah, so I think the whole, and you're right, platformization is a thing now. And I think that's actually almost a maturity of, you know, the move to the cloud, right? Because now people are thinking of distributed systems as systems. This is systems thinking. That's what platform is, right? It's a holistic view of your problem space. Your problem domain, right? You know, people used to do these siloed, specialized products. And the problem is, the problem is no longer siloed.

>> Yeah.

>> The problems, you're a whole attack surface, right? So you, you got to go wide. You have to have an understanding of that domain. And, you know, and one of the other hot topics here, of course, is a generative AI, right? We came back from Barcelona and all the sudden, generative AI. It started stolen the thread, right?

>> We've been joking that it was invented over the holidays. AI.

>> Yeah, absolutely. It is sort of funny cause it's been kicking around for a couple years, but boom, it just exploded, right? I mean, it's sucked all the air out of the room. And, you know, the funny thing, it's, it's great for us, cause, you know, Enterprise Web, given our nature, is the ultimate serverless backend for generative AI, right? And we're already working with the big players here because if you think of it like, this is a revolution in data analytics. That's what it really is about being real time conversational, interactive. Well then you want your backend to be able to support that. If your backend's sort of static, you're not going to optimize against this information flow that's constantly coming in, right? You want your processes to be able to react to what's going on in the real world. You want your transactions to be optimized for that as well. Your operations to be synchronized. So, so we're, we actually believe not, we already have a, an industrial grade, no-code automation demo using generative AI. We just started, I've been here all week in the valley running up and down the, the usual streets, right?

>> Yeah.

>> And it's blowing people away.

>> Dave Vellante: So what are the requirements, Dave, from the backend to support the sort of new wave of generative AI workloads? What, what are the, what are the fundamental characteristics of that backend? What do they have to look like?

>> All right, so, I mean, obviously this goes pretty low, you know, but, so I think starting at the highest level and then working down, because I, I don't want to lose people too quickly.

>> Oh! Our audience,

>> I've been accused, I've been accused of that before. And but at the highest level, you can say this is two sides of the same coin, right? This is just the latest in advance analytics, meaning the latest in advance in automation, right? That's, that's really what it is. And those two have always gone together symbiotically, right? Data drives process, process creates data drive, right? And it's just a loop, right? And that's what you really, in a pure sense, that's what you want. You want to be able to observe the environment, use those observations to drive next actions. You want those next actions to be really fully optimized for what you, you know, based. But you also want them to be safe and secure. And some of those are some of the issues with generative AI right now, right? You want them, you know, setting aside the whole IP issue about what generative AI's trained on, that's one issue totally sort of orthogonal. But the other issue is generative AI is not always right. Sometimes it's adamantly wrong, sometimes it hallucinates. I mean it's, you know, it's working over those large language models. It's essentially drunk on data, right? So you need to put constraints on top. And that's the difference. Again, that's the historical difference. That old Olap versus OLTP, if you remember those days, right? Analytics versus transactions, right? Probabilistic statistical on one side, deterministic rules based on the other side. You could think of those as almost being two sides of the brain, right? Left brain, right brain. You got sort of this creative side that's constantly reacting and then, then you have sort of the logical side of the brain that's actually has models and it's applying rules to things. Enterprise Web is bringing sort of that industrial grade side to it. If you look at a lot of the automation demos for generative AI, you, they're clearly experiments really like rough beta prototypes. They even say it when they're doing it. And they're sort of like, if this and that kind of stuff really stuff. And they're not thinking about security and they're not thinking about IT governance. Well guess what? Nobody at the enterprise is ever going to touch that, right? You know, the enterprise needs, I mean, this is, you can't have creativity in your transactions, right? You can't, right? It's like, yeah.

>> You don't want yourself,

>> It's got to be binary.

>> Yeah, you like your banking ledger to be, you know, hey, you know, it's like put in a hundred, it shows the a hundred, right? You, you, you want that kind of transactional control. And so Enterprise Web is adding that, let's say rigor, right? So, so in this demo that we have today already, we can demonstrate it working, it's great. We're essentially, it's a Microsoft stack we're working through today for the obvious reasons, it's really the one that's breaking away. But we're working through Jarvis, which is our NLP, right? So we can actually talk to Enterprise Web Jarvis through open AI, right? To enterprise web. There's another product in between I'll get into when I go into the technical details that you ask. But essentially I can talk to Enterprise Web and say, I want a network service, something really advanced, right? That's even hard for an engineer. I want this, please compose this for me. Then when it composes it, I want it to be optimized this way, configured this way, and then deploy it on the network. And this, in a couple of minutes, I could dictate this to the machine. The machine does it and then it shows me the state, right? So I can compose, deploy, and manage on an ongoing basis the day zero one and two working with generative AI.

>> Dave Vellante: So you can say, okay, optimize for low latency and give me an estimate as to what it's going to cost in a monthly.

>> Dave Duggal: Yes.

>> Okay, that's a little too expensive. So dial down, dial up the latency a little bit, and you can speak to it.

>> Absolutely.

>> That's your language way. And then optimize your infrastructure for an outcome.

>> Yeah. And energy, and this is exactly, you know, in telecom they're calling that intent, right? [Dave Vellante] Yes.

>> Managing the declarative intent, right? So a developer can say, you know, energy efficiency, maybe it's not latency, maybe it's energy efficiency, right? That's my, so manage, I don't want to care how you manage energy efficiency. The details of that are way too low level. You know, it's down to the network, it's down to the, the RAM, the cores that you're touching a lot of things at that point. So, but you know, does a business really care? No, it just wants its SLAs and SLOs met, it wants its intent met. Honestly, if you think about we've spent the last 10, 20, 30 years of computing thinking so much about our tools, right? Completely independent of the business behavior and objectives and now we're actually almost at that stage, right? That inflection point where we're really going to enter the 21st century for the first time, right? I'd argue that we haven't been in it, right? We're still doing 20th century automation now. We're also going to be in, the focus will be you talk to the machine, you express your intent and the system in onset and it's fully traceable, right? It's fully governed.

>> Talk about the self-healing aspect of, I noticed that was in your LinkedIn post.

>> Yeah, yeah.

>> Because that's the, that third wave of gen ai. You have prompt engineering, prompt ops, prompt tuning, which is, let's see the query operationalizing it.

>> Yeah.

>> And then self tuning, tuning it, which means, yeah, it's on its own. Where do you guys fit in on that piece? Cause I can imagine that's where, when you give an order into the command provision, this, the hallucination side's a concern. So I'm assuming you're doing that self tuning, self-healing. Take me through that piece of it.

>> So, so we do this, so, you know, we're using the Microsoft Gen AI in front, right? So that's the, that's, let's call that the, the Jarvis and the Open AI up front. We put another technology between us. Essentially what you want is a vector-based time series database. So this goes to, to your question actually the both of your questions, because we do this cause we want an intermediary, we want to be arm's length, we want a nice separation of concerns between us and the generative AI for exactly the reason you just described, right? I'd want that hallucination to flow into my commands, right? So what what happens is the hallucination is working through this time series database, which is putting that data in order, right? It's taking this massive stream of event information and we're, it's applying rules to it and then just throwing events to enterprise web, right? But those events are already sort of filtered in a way that Enterprise Web can respond to them. And what it's doing is that, and this we work with a company called KX, which a really interesting company, you guys should have 'em on some time actually.

>> What's the company?

>> KX

>> KX, kx.com, 20 of the top 20 realtime.

>> They're AI ML partner, right?

>> Yeah, yeah, exactly. Yeah, so 20 of the top 20 financial institutions use 'em for realtime training, right? They wanted to enter telco. So they actually partnered with us because we're, we have tech telco domain expertise, right? In this case, so now totally different use cases and more, cause General Enterprise Web is generalized, generative AI is generalized, KX is generalized for analytics and everything like that. So we use 'em as an intermediate. So what happens is they're, we're arms length from generative AI when it's talking to KX, KX is then translating it into the symbolic language is semantics that we speak for us, right? And it's almost like a universal translator, right? Translating back and forth between Vector that.

>> Brings coherence.

>> Coherence, right? Because essentially what we're allowing it to do is the, through the intermediary generative AI is walking our graphs, it's walking our domain models and introspecting our catalogs. So when I actually, when it, when I call, talk to my system and I say I want something, it's literally walking the enterprise web graph saying, oh, I found it. Oh, and here's the object I want to use. It's a Cisco router, a Juniper router, or whatever it is. I'll compose this with this and this and this, and maybe it's a fort net security firewall, you know, whatever, whatever it is. It'll put that together, apply my constraints and then implement it. The controls come on the Enterprise website. Completely discreet. So generative AI is enabling the conversation that's super powerful. Almost think of it as the most advanced user interface you can imagine, right? It's now Star Trek, right? This is, you know, all right.

>> We said in theCUBE one time, everything in Star Trek will be invented except for transporter room and that automatic food making.

>> Yeah.

>> But you just gave an example of whatever, a Cisco router or, or Palo Alto firewall, whatever. What is it that allows you to not care? And then is that different in telco?

>> So it, it, it's interesting. So part of it's what we've already had the sort of core enterprise web platform in enterprise web, essentially we have what's called an upper ontology, right? Upper ontology just means it's the, the generic set of concepts and types that runs across all enterprise businesses, verticals, whatever. Every business has people, it has organizational units, it has facilities, right? Those kind of things. And then, then of course, from a cloud and distributed systems perspective perspective, there are common types, right? There's formats, there's protocols at the upper level, protocol enterprise Web maintains that. It's sort of the universal concepts that apply to everybody. Then Enterprise Web allows you to model a domain, like whether it's telco, whether it's life sciences, whether it's IOT, it doesn't really matter to Enterprise Web. We enabled you to rapidly model a, a complex domain. And then what happens is you're onboarding objects into that domain, right? So that router you were talking about, that firewall you were talking about, right? So what you're doing is, instead of doing that old kind of point to point integration you used to do in the stack, what you're doing is actually just mapping up to those graphs. You're just mapping in metadata. So we're saying, hey, we understand Juniper Router or Cisco Router or Palo Alto Security firewall or Fort Net. We understand them completely in metadata. Enterprise web is a hundred percent configured by metadata, right? All enterprise web is a graph model, which has policies, which has metadata policies, you know, and types. And then it's stateless functions. And Enterprise Web uses that metadata to efficiently hydrate context for stateless functions. That alone is a super big idea cause serverless.

>> You have a very big idea. So the super cloud aspect of the telco thing, is super impressive. What you're getting at now is the modern platform vision. It's, enterprise is hard. We had Jeremy Burton on, who was a very distinguished executive over the years, now he's doing an observability startup. He's like, it's hard to do a startup, especially in the enterprise, takes years. Where are you guys at?

>> Tell me about it.

>> Take us through your progression because you guys have a solution now. How does someone get in? How do they get involved? What's the deployment? How do they get, how do they adopt the technology?

>> Yep.

>> Cause it is definitely in line with what people want.

>> Yep.

>> Some people don't have the skills. How do you guys do the managed services?

>> How does someone engage with you guys?

>> Yep.

>> And, and and how long have you been around?

>> Yeah. Fantastic. So, so I've been,

I've been personally? >> No, I've been around for a long time.

>> No, Enterprise Web.

>> So actually, so I had, I had started turnaround growing several companies in a row. And, and so around the time of 2008, 2009, you know, pre, pre before cloud took off, cloud doesn't really take off till 2012, before Kubernetes, before microservices, right, before containers. My previous engagements, my roles were, I was like seeing that the 20th century automation tools are getting the way, the kind of things I wanted to do already, just from a business intuitive perspective is like, why can't I do, why does it tell me what I can't do? I'm the business guy, I want this done. Why is being, you know, when I asked some behavior to be implemented, I don't say do that in concrete. Right? I'd like it to still be adaptable if my, you know, if all of a sudden my needs change the, the competitive landscape. So I start, I, in 2007,8, 7, 8 I read three, 400 academic papers and industry papers. I seriously, I actually, I left my prior job and I said, you know, I'm going to take off. I'm going to do sabbatical, I'm going to figure this out. And so I went out and just did all this research and just look it up, like what, what, why, why can't people be more dynamic? And basically it was the old, this old software was based on the old constraints of hardware and infrastructure. And when the hardware and infrastructure kept on advancing the software didn't.

>> Yeah, that's.

>> Right. You know, partially because those models of selling packaged software were pretty profitable. Selling lots of different components.

>> Highly motivated to keep it the way it is.

>> Yeah, they didn't want to capitalize it, right? So I said, you know, actually this should be horizontal. This is really capabilities architecture, right? I should be able to say, I've got a catalog with my capabilities and I have business logic over the top. My middleware should be this thin horizontally architected layer between my business logic and my things.

>> Yeah. >> And that's all that matters. I don't want to care how it deploys like we talked about earlier. So I started doing that and essentially I just work behind the scenes. I just self-funded it. I, I've taken no money. I've bootstrapped this whole company this whole way to profitability, right?

>> Congratulations, that's awesome.

>> Thank you. Awesome, yeah. Not an easy road, I don't recommend it to everybody.

>> It is, well it is, is an elite, it is an elite title to have that self-funded start.

>> Now you have options like you that, that you wouldn't have had if you done,

you'd taken money >> Well done.

early on right . >> Thank you. So we're in production, some of the world's largest companies and we go to market new channels, so.

>> What kind of company, how big are they?

>> Well, Red Hat, so last, on our last show we had Red Hat, you know, Azhar Sayeed from Red Hat was on with us. And that was a great show, so Red Hat looks at us as, you know, maybe being the application layer over OpenShift right. We have partners with Intel. We just did that award-winning test bed, five grand. We're, we're the closing loop with our partner AKX on the AI ML side, closing the loop from reported event to mediation.

>> Yeah.

>> In 300 to 400 microseconds, not milliseconds. The state of the art orchestration is 11 milliseconds. Alright. I'm at three, right. I'm orders of magnitude more performant they are. And the reason is I have no cruft, right. There's nothing but a graph and the stateless functions once

something happens. >> Yeah. We flatten that graph.

>> You got to love the land, >> Execute the functions.

>> Got to love them. Got to love serverless.

>> Yeah, oh, oh, and we implemented it. So we have our own, so we have 19 water patents, multiple patents pending, right so we actually, cause we existed before, there's other technologies because you have to remember.

>> At least she's throwing money at you right now.

>> We're getting a lot of interesting conversations. And the generative AI thing is, without me saying anything, the biggest players, and you know, there's only two or three, the biggest players in generative AI are super interested because right now every organization in the world is reconsidering their operations in light of generative AI.

>> Totally. Right. In the last two months, you.

>> If you're not, you better be.

>> And their concern is about the hallucinations which you're trying to address. That's the opposite of like, I don't, I want pure, reliable, scalable.

>> Deterministic.

>> And it's got to be portable.

>> Low latency, high performance, portable. Right, we're 50 megabytes, you know, again, we got all these great characteristics because we rethought it, right. And people didn't think it was possible so they didn't try. And also I think one benefit.

>> Well I think the architecture you laid out also business apps essentially could be put cloud right there. So you actually had, the architecture was perfectly lined with how the cloud spun.

>> Yeah.

>> Hence our super cloud and our rendezvous. Our serendipity meeting you.

>> Yeah, yeah.

Exactly. >> That's awesome.

>> Yeah, well actually, even before we went on on theCUBE with you, I went on theCUBE with you. You, you were already talking about Super Cloud Paws. So I saw that, I was like, ah, I got to follow up with

these guys. >> Yeah, yeah. We're in the same tribe, let's go.

>> These are my peeps. >> Yeah. Yeah, well I'd been on theCUBE back in 2015 in a structure conference with George Gilbert.

>> I remember that.

>> Back in the day.

>> I do. You still working with George?

>> Yeah, actually I've been having some conversations with him.

>> He's a great guy >> And he of course dodged me.

>> I remember Digital, digital Ocean back then. But they, I interviewed those guys too.

>> Okay, yeah.

>> Yeah, it was good time.

>> Yeah, so, so we go through market, through channel. So we have SI people like Tech Mahindras, one of our SI's you know, Red Hat, Intel, you know, Amazon.

>> How does someone engage with your customer? What, who wants to engage with Enterprise Web? What do they do?

>> Well so I mean, anybody can,

>> So a channel partner or,

>> Yeah, so they can inter, they can go directly to any of those channel partners or they can obviously David enterprise web dot com,

can I do that? >> Yeah.

>> Am I allowed to do that?

>> Of course, yes.

>> David enterprise web dot com, so they, anybody can reach out to me and, and I'll direct them to the right partners that can do service delivery and sport. I'm focused on building the world's greatest software company because I think the world needs it. Every epoch, right.

Every period >> Yeah.

>> Requires that new IT company to come and do that, and, you know, maybe it's Hubris, right? I'm a guy who sits in Glens Falls, New York and mostly working my t-shirt and jeans. But you know what, actually every once in a while the world, the ocean does need to be boiled cause it's rancid, right. And people need to rethink things.

>> Bill Gates start was in Seattle. No one thought anything about Seattle except,

same with Jeff Bezos. >> Yeah.

>> We're going to put Glens Falls New York on the map.

>> I love it.

>> And so, yeah. So it, it's been a journey, you know, it's been a lot of fun, I mean, I've learned a ton, right? I've become a student of organizations, you know, working with all these big players. I, I've been squashed so many times I've, I've heard every naysayer, but you know what? I knew our ideas were right.

>> Stayed the course.

>> Yeah, and, and that, you know, you know, luckily we had enough friends, partners, customers to, you know, support us. We knew that we were doing things right. We were validating it in the background. And that's what I knew I had to do. When you make big claims, you need big validations.

>> Yeah. >> So that, that's why Well, you got customers, you got this,

they're in production. >> Yeah, yeah. Big customers.

>> And your next steps now is what? Scaling to go to market,

>> Scale out. Scale out, this is ready to rip, right. This, this is, this has been proven and,

>> And you need funding to do that.

>> Yes. So yeah, we're at that scale. So we're at that inflection point. And so,

>> So growth capital, you're ready to go.

>> Yeah and we have other conversations too, but yes.

>> And you have, yeah or right, okay, but, and you have options because you,

you own the whole thing. >> Yep.

>> That's awesome, congratulations.

>> Thank you very much.

>> The gen AI piece is beautiful, so congratulations on that.

>> The, the gen AI almost, you know what the thing is partially, I think is now this is really our moment. I think we were on the right track anyway, but middleware's just never been sexy, right? It's been, it's plumbing. So selling middleware is a thankless job, right? And, and if we do our job well, it's like a utility, you shouldn't even know it's there. Gen AI makes us super sexy, you know,

being the back end to Gen AI >> Yeah. is a, we're talking about what, a hundred billion trillion market opportunity.

>> Yeah. Well we talk about it all the time that Amazon turned the data center into, into an API. And generative AI is going to turn the way in which we interact with technology into language

>> And it's automation compatible. People know what automation is on DevOps and operations. They see value in automating undifferentiated tasks. So I think it's, it'll, I think it'll play well. And then then having this more self-service, provisioning,

self-healing. >> Yeah.

>> All right guys, we got to go. We're, Dave always pleasure to see you.

>> Fantastic.

>> Great seeing you come out.

>> Great seeing you, keep right there. We'll be back with our next guest from RSA 2023. You're watching theCUBE from Moscone West in San Francisco. Right back.

>> Okay.