 Live from New York City, it's theCUBE at Big Data NYC 2014. Brought to you by headline sponsor, Juan Disco, with support from EMC, Mark Logic and TerraData. With hosts, Dave Vellante and Jeff Kelly. Welcome back to New York City, everybody. And we are really pleased. We're going to dig into the networking space a little bit. Dave Husak is here. Founder and CEO of Plexi, a company in New Hampshire. And networking is a space that's obviously been dominated by Cisco. The company's got two thirds of the market and has for years. Networking's this big, top-down, hierarchical animal. Compute scales out. Storage scales out. I think Plexi is the vision of making networking scale out. So, David, welcome to theCUBE. Thanks for coming on. Thank you, Dave. I'm glad to be here. So what's the story, right? Like I said, compute, storage. Everything's going to scale out. The network has to go scale out, does it not? It does. It's the element that makes all those other things work. It's for high performance, scale out data, accessibility, and computing. And Hadoop is obviously a fantastic use case, a fantastic expression of all that. And networking is central to making that work. So I want to get into that. But before we do, tell us a little bit about Plexi, how and why you started the company and give us a little background on it. Sure, sure. I've been a networking guy my whole career. Way back when in the 90s, I worked at one of the companies that actually built and installed the first switches, ethernet switches in the world, before we even need to call them switches, right? And... What did you call them back then? Well, actually, Ethernet Express is what our product was called. Back when, you know, it was 10 megabit ethernet and things were connected after the iBackbones and stuff. That was megabit, folks. Yeah, megabits, megabits. Electrons were way slower back then. And so at some level, I helped create the problem that we're now solving here at Plexi. That's what entrepreneurs do. Yeah, it keeps the cycle rolling. But we have basically, with Plexi, the inspiration was actually a lot of what you, the problems that you set up already is that recognizing that the way we've been building switch networks, you know, hierarchically for the last 20 years now, doesn't scale the same way applications and data is scaling out. The requirements have changed. And, you know, 20 years is a good run for a technology and it was time for something different. And a lot of it, so the inspiration for Plexi, we got started with the idea sort of germinated and developed back in 2010, raised our A-round in 2011, followed that up with a couple other, you know, funding rounds and rapid succession. And, you know, we think we have a take on networking for scale-out applications and Hadoop and Big Data and that offers, you know, really compelling advantages compared to the alternative, which not only is Cisco, but is kind of everybody else, right? They all essentially build fundamentally the same kind of product, whether it's leaf spine, three traditional Cisco three tiered course, switches and things, right? They all have this property that, you know, if you want to build wider, you have to build taller, right? And so it's fundamental geometry. You can't get away from that, right? We have a different way of wiring networks. We have a different way of orchestrating networks. We employ SDN as the methodology that glues all that together and delivers a level of automation and visibility and manageability that the other guys can't touch. So I'm envisioning this sort of mesh, this fabric that's sort of controlled by software. I want to get into the Big Data and then we can maybe get into your technology a little bit as well. I remember at Hadoop World in 2010, I didn't know much about Hadoop. I'd read a little bit about it, but when we started to talk to some of the practitioners, Abhimeda in particular, who's the guy we're going to have at our Capital Markets event tonight, said, listen, the fundamental is you ship function, you ship five megabytes of code and not a petabyte of data. And so I said, okay, well, it means tons of data. I get that. It means getting compute closer to the data. I get that. What does it mean for networking, that fundamental notion of distributed computing, function shipping, what does it mean for the network? It's all about agility. It's all about being able to place the computing with respect to the data, with respect to the resources it's going to consume in a way that ensures that it's making, it has access to the data it needs. The data is replicated across the right kind of boundaries to ensure that the reliability and resiliency is there. It's ensuring that one of the things people don't talk about enough because it is a hugely important problem, especially as Hadoop and Big Data application sort of proliferate. But security concerns, right? In some cases, payment card, healthcare, there's a lot of applications out there where there are regulatory requirements that are problems that have been solved in sort of conventional data file systems and processing architectures, but also need to be enforced in big data installations as well. All of these things are mediated historically by big boxes of network functions and firewalls and filters and such that are static, right? Moats that are dug around the castle. Exactly, exactly. And that's sort of counter purpose to the way we kind of now think about employing computing and accessing data in large sets that may be geographically distributed. Maybe my data's on my data, some of your data to actually accomplish a particular application. So the thing that ties all that together can't be static. It has to be agile. It has to be fluid. And networks need to evolve out of sort of a rigorous hierarchy that's defined. Networks of today are effectively defined by where the wires go, right? That's no longer optimal. It's no longer practical. Okay, so everything is virtualized and fluid on top of it. Right, okay, right. It becomes a mess and really from a security standpoint, very hard. So how do you solve that problem, number one? And number two, how do you convince the Cisco certified engineer that they should take a chance on Plexi? Well, hopefully we're making it less of a chance as the product, we are delivering our sort of second version of our products with like a lot of storage. So we're a little over three years old, right? And we now have delivered Rev2, which is, as you know, for startups, Rev2 is the real first product. The Rev1 beta products where the test the water kind of things, right? But first, yeah, it is less of a chance because we are fully interoperable. We've plugged into lots of people's networks, lots of environments. And so the ability to add Plexi, Plexi Agile Networking capability to existing networks is easy. I did actually worry at the beginning about the question you posed, right? Like, because, you know, back then we were telling the sort of the big story, the, you know, what it's gonna be like to have a network that works this way, that software defined and Agile, and of course that narrative is immediately appealing to the CIOs, the application architects, the guys who, you know, the users of the network, as opposed to the guys who build and operate the network, right? And I did worry at the beginning about, well, you know, we definitely have to engage the networking engineering community, you know, the guys with the CCIEs and JNC, all those guys, and, you know, they're gonna be resistant to change because, you know, they've been doing, a lot of people have grown up in a career where they'd never known anything else. It's been 20 years since we've been basically doing things the same way, right? And they have huge investments in scripts and CLIs, you know, the kind of the archaic ways that networks are built and operated generally today, right? It's their house. It is their house. And so you can't, you know, not in my house, right? You can't just expect them to, you know, to accept something new and different. What I didn't, so we started engaging, you know, like about, you know, as we're delivering our beta products into our first customers, you know, we were, I've obviously had to engage, right? And what was striking was how much of the frustration on the network user side was actually mirrored in the network engineering side, right? These guys knew they were behind. They knew that the guy who was orchestrating, you know, vSphere and vCenter was, you know, can make thousands of servers, you know, spring into existence, vanish, copy themselves, dance across the networks, you know, with keystrokes and mouse clicks all day. And they were the guys that were still, you know, typing cryptic commands into CLIs at 3M on Sunday mornings, right? Which was, you know, the gap between that is just crazy, right? And they knew they were behind. They knew they didn't have the tools. They weren't getting the tools from their vendors that could allow them to keep up. So you're saying they were more receptive to it? They are more receptive. I think the, you know, as we refined our proposition and matured our products, it's now becoming, like I said, hopefully it's not as even scary anymore. It's actually, the advantage is outweigh the, you know, the risk of bringing it a start. Well, it helps too if you're 10x better. Are you 10x better? We are more than, you know, depending on all the axes, you know, we are more than, it's just a different model. It's, in fact, it's, you can do things with Plexi that you simply can't do with other guys. So that's infinitely better. David, we got to leave it there. Get in the side. Thanks very much for coming on. Thank you. Really interesting. And we'll be following you. Hope to see you back into England. Sweet. All right, keep it right there, everybody. We'll be back right after this word from Big Data NYC. This is theCUBE.