 From the Julia Morgan Ballroom in San Francisco, extracting the signal from the noise, it's the Cube, covering Structure 2015. Now your host, George Gilbert. And we're back live. This is George Gilbert. We're at the Julia Morgan Ballroom in downtown San Francisco at Structure 2015. We have another special guest with us, Dave Dougal of Enterprise Web. It's kind of topical that Dave's following on Derek Harris, who was at Mesosphere, because we have something that's somewhat analogous. Dave, welcome. Thank you. Thanks for having me. Nice to meet you. Good to have you. Cheers. So tell us. A lot of people are more familiar with Mesosphere and where it fits into the sort of technology stack. Tell us, maybe in terms of that, how you're similar and what additional functionality you have. Great. So I think you hit the nail on the head when you said the stack. And if we look at the 21st century, and the world is flat, the cloud is horizontal, our APIs and services are loosely coupled, but we're still building stacks, Kubernetes, Mesos. We have cloud stacks, big data stacks, we have web app stacks. We're building all these stacks, and then we wonder why our data is still siloed. We're essentially rebuilding our silos in the cloud. And let's not forget, some of these technologies are exceptionally hard for developers to deploy and maintain. So where Enterprise Web took a really different approach was when we formed the company in 2009, I was looking at the technology, IT, instead of being a service to me, was becoming a rate limiter to business. And I wanted technology to be more flexible from the business down, because all that really matters, because let's be honest, the business doesn't really love IT. IT is infrastructure, it's a resource, it's a utility. It should be in the background. When you pick up your phone, you get a dial tone. And if it wasn't there, you'd be really surprised, right? It's a utility to you, right? So the fact that we talk about all these stacks and these technologies and different code languages is actually a sign of the immaturity of the marketplace. Actually, because all that really matters is application behavior. So Enterprise Web is an idea that suggests that in the 21st century, the middle tier should be a horizontal layer, horizontally architected, elastically scalable, lightweight layer between the behavior I want, the application logic, and the objects I have access to. Whether they are nodes of compute storage or network, or whether they're policies, app models, process models, data schemas, whatever they are. They can all be represented as objects in a singular fabric, and that's really the notion of Enterprise Web. Okay, so let's boil that down now to examples that some of our viewers can relate to. You had mentioned to me earlier, a sort of a systems integration web. Like spell out the scenario, the sort of the steps that you would build in that pipeline. Okay, great. So if you think of what a pipeline is, a pipeline is essentially a decomposition of a process, right? I mean, it's a functional decomposition, right? Each step of a pipeline adds value, but usually what we're doing is to take something from a source to a target, right? And usually have several intermediary steps. The problem is for the last 20, 30 years, the way we affect a pipeline is we actually have middleware components, actual physical boxes that actually support each step in that pipeline, right? That creates a lot of indirection, a lot of systems complexity. It causes latency, it also causes switching, context switching, it has an expense, right? What we're suggesting is that really what we want to do is all those transformations in a single layer, right? And eliminate that stack architecture, right? Because that stack is non differentiating, right? No, but nobody in the business cares who your ESP is. And would it be fair to say that even though we have a mental model of stack being south to north, or north to south, this is a stack that sort of- East to west. Yeah, it's horizontal, absolutely. Because all those steps are essentially services as opposed to components, right? So it's still a decomposition, but instead of having a physical component actually representing each processing step, I actually have services doing it. And now I'm actually in a cloud native kind of model, as opposed to sort of having this impedance mismatch where my cloud and everything else is horizontal, but my app is still vertical. So tell us about what some of those services are and what they were as a north, south, or east, west stack. So here at Structure, we did a workshop with PwC. I was really pleased. Glenn Hobbs, managing director of strategic technologies for PwC, did a presentation with us where they got up and talked about their use cases, which was great. So they had five people take six weeks. That's 150 man days. And they got a pipeline from Salesforce to Tableau, which is going to be pretty common or anything like that, a CRM system to a BI kind of system, right? With multiple intermediary steps, including Hive and other technologies. And they took 150 man days and they got down to seven minutes. Actually, six minutes, 53 seconds, right? Then they gave us and they thought, hey, maybe we can do this without the stack. They knew us, they were exploring different kind of use cases with us. Let's try to put Enterprise Web on this. We took one man four hours. We get it down to 300 milliseconds, all right? We stripped out the cruft of that network where they literally said, they wired together a stack. And remember, they had to wire that stack together before they even built their solution. And remember, that's also going to be very brittle too, because any change in any of those components will break their app. Where in our system, you're just stringing together a set of series in a loosely coupled, restful way, a series of services. And they all dynamically bind. And that's how we get to sort of phase change orders of magnitude improvement. That's, yeah, when you mentioned orders of magnitude improvement, especially in an east-west stack where latency is the biggest issue facing, you know, analytic data pipelines as we move to an era of analytics. So how does PWCC applying your technology to their, you know, future solutions? Sure. I can't really speak for PWC directly. I think they kill me, but. Anonymize that. Yes, okay. I think part, so we go to market through partners. Yeah. And partners look at us as a vehicle to bring transformation, highly differentiated solutions to market. And that's our niche, right? That's what we want to do. We could, our platform is really a general purpose. It could coordinate anything. We're doing industrial internet, as we discussed earlier. We're doing, we built a fabric over SAP connecting its components, which are notoriously hard to connect. We, we, even for SAP, even for SAP, right? People spend millions of dollars per year just keeping the integration between their SAP components. But in addition, we also dynamically construct UI. So we have some of the world's largest hospitals running enterprise web for their entire operations. You know, billion-dollar organizations using us to build UIs for complex human processes and not very compliance-centric processes. That's a very diverse set of projects. Because we're a generalized platform. So, okay. Help us set some context. Perhaps no 100% direct competitors. But who, who's trying to solve similar problems? Yeah, and, you know, and I've spoken to a lot, you know, I know a lot of the speakers here. David Lynn, think I'm just wrote a white paper on us. And, you know, I know, I met Adrian at Cockroft out here at a structure event several years ago. So I get feedback about who's doing related things. And I think there are a lot of people trying to raise abstractions, right? They're trying to say, you know, we can't be in the CLI anymore, right? We can't be, we're too much, we're just like, we fell out of the networking cabinet. We're still wiring things. And the analogy I like to say is that, like, you know, at a time we used to have switchboard operators, right? And if we, if we still had switchboard operators, we wouldn't have this global communications infrastructure that makes the cube possible. So we had to, we had to get rid of the man in the middle to facilitate those real-time dynamic communications. That's what we do. We use intelligent agents to essentially interpret every request to our system against that object space that we talk about, against that fabric and compute a real-time personalized response. If I didn't know better, I would have called that like a object request broker that was, you know, au courant about a couple decades ago. Well, you know, so Richard Soli, the head of the object management group, Richard Soli, you know, from the OMG, which owns the Corba standard and the BPMN standard and the UML standard, he's actually one of our advisors. So we've been surrounded and lucky to have attracted the interest of a lot of powerful people like that. So I think actually a lot of the old ideas are resurfacing. We'd like to think we're tied to first principles. We'd like to think we're tied to like Malone and coordination theory and Strayke and Milner and like the, and even Turing, right? Because these guys, even before the systems existed, before even the infrastructure could have supported it, they did it all in their head as math, right? They understood this as graphs and they understood the right way to do things. And now what we have now in 2015 is the infrastructure can support us. So what we're doing is just very efficient graph processing of very complex objects but you're completely spot on. So these ideas have common waves in the past. And now the infrastructure can support it. Yeah, so the abstractions were limited for the infrastructure of the time. They throttled themselves because that's what the technology could support. Now we're saying, hey, gosh, we live in a world without constraints, right? Compute storage and network are unlimited, right? I can just turn up and scale that up and down. So it's a different world but honestly some of the best ideas are the old ones. And I think actually when you talk to the innovators, you're actually hearing the echoes of giants actually. Right, yeah. All right, Dave Dujo, we have to leave it there and very much appreciate your joining us. It's a great meeting, George. This is George Gilbert. We're at Structure 2015, Julia Morgan Ballroom and we'll be back in a few moments.