 Hi everybody, this is Dave Vellante. Welcome to CUBE Conversations. Within the Wikibon community there's a lot of discussion around replicating public cloud capabilities, particularly scale and agility inside of the private cloud. I'm pleased to have Rich Napolitano here. Rich is the CEO of Plexi, long time friend Rich. Welcome back to theCUBE, it's good to see you again. Great to be here. It's a pleasure. So you guys have had a busy year so far. It's only opening day in April and here we are. You guys have new funding announcement, new product announcement, give us the update. Now again, thank you, it's great to be here. So just for the record for everybody outside of the East Coast, it snowed last night and so it was a pretty snowy day on April 5th, opening day for the Red Sox, so Go Sox. They plowed the field this morning so they're out there right now. But we're hardy here in New England so that doesn't stop us and Plexi's continued to march on. In January we announced a new investor, Google and Google's been following the company for some time actually and excuse me, we announced the investment in January and if you talk to the Google investment people, they would tell you that it's really a statement of architecture that Plexi really has the right network architecture for the future, really for the third platform. Our ability to scale, our ability to be software defined, our ability to shape the infrastructure around applications in data is what they fundamentally saw. So when Google, how much you can say, when Google makes such investments, it's looking at how it can apply such technologies to its own business, is that right? Or is it more looking as sort of an investment vehicle or how do we approach it? I mean, well, it's Google Ventures so they're interested in making money. It's no accident. There's certainly an opportunity inside the company that wasn't the main impetus for it. They want to shape the ecosystem of the world to be more fluid and more cloud-like whether it be public or private. And that just creates an ecosystem which they can monetize in many different ways, including from the investment perspective. Okay, so also big announcement. You've announced what you call the industry's first programmable network fabric for converged infrastructure and storage environments. Let's unpack that. What does that all mean? Yeah, so I mean, we've learned a lot. My around 18 months tenure now with the company and last year was our first year of significant revenue growth and we're very pleased with those results and expecting even bigger year this year but we now know really the nature of the customer and the nature of the customer are these cloud builders and they're building clouds. And in particular, they're building either the service provider customers that we have are building big service-oriented networks or private clouds. And so what people are trying to do is create agile, dynamic, flexible private cloud infrastructures that give them the same agility they would have in the public alternatives. And so Plexi represents a technology, a building block that allows them to build these private clouds that are as agile and flexible as public. So a lot of people look at cloud as a place to store data or run apps as a destination, if you will. What we see is really the business impact of cloud is about the operating model. How are you and your customers affecting the operating model in a cloud context? That's a really great question. I think really the essence of the strongest part of our value propositions. We have a CAPEX savings. If you look at it on our architecture, you have 90% less cables, one third less devices. But really the big saving comes from the operating expense because our network is really invisible to the application, invisible to the operating environment. So we're so well integrated into the VMware environment that the network is really shaped around the VMware cluster. Or we're so well integrated into, very quickly as a Fadoop, Hortonworks is part of our announcement, Cloud Hour as well, where we can shape the network and therefore compute and other resources around the application and data. So automatically, which means you're not provisioning networks in bandwidth and trying to restrict flows at every client in a network to ensure your overall SLA. The network is shaping itself dynamically through software control, automating these manual processes through DevOps, through, you know, think about, I mean, in essence what we are, our founder says this a lot, Dave Hussack, we're an API driven network. You can program our network to behave differently, to be shaped around your data and your applications as a huge opposite. So this really is all about the infrastructure as code theme. I know you like data, so we just put out a report on, we call it the true private cloud, kind of tongue in cheek because a lot of the private clouds out there, we call them cloud washed. But we quantify that there's about, over the next 10 years, about $200 billion that is gonna shift from non-differentiated, you know, infrastructure management into vendor R&D. That's really your business. That is our business, certainly. Okay, so talk about that a little bit in terms of how that's affecting organizations in terms of their skill sets and how they should think about moving forward. They're, you know, in fact, I mean, it's interesting that you bring this up. We know we have a really hot prospect when they latch onto our highest level of our value proposition, which is really DevOps. When they start understanding that they can program our network seamlessly, when they can manipulate our network through software to meet their overall SLAs, that's when we win. We just, we don't need to say much else. They almost self-identify, but they self-actualize in terms of really understanding the value proposition. So that's, and that, everyone is trying to automate. And so, because they need to take these manual steps. Precisely, our network is not managed as a group of individual devices. As individual devices, our network is managed as a network. So when you program our network, our APIs work across the entire network. So if you set up a VLAN between California and Tokyo, you're not touching the hundreds of devices in the network. You're just telling the network that you want a VLAN from here to here. That has huge implications because the alternative is touching hundreds of potentially devices in your network and programming through a CLI or typing into a CLI, which is extremely error prone. And some of the kind of traditional networking players will create kind of an overlay script that overlays the CLIs on all these devices. And they think that that's automation. That's not automation, right? Our metadata for our network is fully distributed and is accessible to our APIs. So we've really elevated the conversation up away from the device to the network as an entity that is programming. And that's your kind of your wise to enable that new operating model that we were just talking about. Okay, so now talk a little bit more about the what. What is it that you guys actually deliver? So we build networks. And layer one, two, three networks, it's manifested in hardware, either white boxes or our own custom hardware. And through the right use case you use, they have different properties and different price points. So, but really we're a software company. We have just a couple of hardware people that do a lot of our hardware, some of it ODM from other manufacturers. Our core IP is in two places, Plexi Control and Plexi Connect. Plexi Control is our SDN, it's the only time we use that word, we generally don't use that word, SDN Controller, which allows us to manage the entire network as one entity. Also it does, it gets these requirements for your network from Plexi Connect, which we'll talk about in a second, and then fit and renders these topologies to allow you to adapt the network around your data and your applications. So all of a sudden now compute and storage resources become fluid underneath this network. So you can move them around very, very easily. And the thing that drives that is Plexi Connect and their integrations into these operating environments. So VMware, OpenStack, Cloudera, Hortonworks, you know, other operating environments, Mezos, even a whole sweep of dev-up tools. We have an integration even into Slack, which that whole community is evolving very, very quickly around a set of industry standards. That is our way into our product. So you can Python, RESTful interfaces, leveraging industry standards into us that we feed into our controller then we can fit and render these topologies and manifest them on these physical networks. So talk a little more about the business problem that you're attacking today, particularly as it relates to network traffic, how network traffic has changed. Everybody talks about North-South versus East-West, but you're going a step further. Can you unpack that a little bit more? Sure, I mean, the driving tendon, this is again a lot of the revelation that we've had the last 12 months, is that when we think about people transitioning from platform two, client server, hierarchical infrastructure, right? Clients, a LAN, server, a SAN and storage. That's the hierarchy we've been living in in platform two. And the nature of those networks was heavily North-South. As people have transitioned now into platform three, they're heavily focusing on converged infrastructure. And so the converged infrastructure, one of the first things that modern converged infrastructure does is they remove the fiber channel SAN. And the moment they take the fiber channel SAN, they move all that traffic that used to be on the SAN and in the controllers and put it in an under IP network. And we're seeing a growing problem in these networks because suddenly the networks that were designed 20 years ago for North-South's traffic from client server era are now being populated with a huge amount of East-West traffic, eight to 10 times the amount of traffic. And also now the metadata that's used to manipulate these storage subsystems now is on that network, which is extremely latency sensitive. And the integrity of that data must be protected. So that determines the performance of your entire system now is determined by your network. As a storage guy for a long time, as you know, we solved the latency problem in storage with flash. But now we've just created this new set of problems which is the moment we converged infrastructure and removed our SAN, now the network is the next new bottleneck. And we need networks that can adapt. So our integrations for, you know, V-SAN or ScaleIO or Nutanix or SimpliVity and others allow us to dynamically shape the network to allow to avoid congestion. So all different types of traffic in North-South, East-West, metadata management, recovery traffic for back in storage subsystem, the various stages of Hadoop processing can be separated on a network and avoid these hotspots that we call micro, you know, microburst in a network where you have a hotspot in your network that you really can't see because you think you have enough bandwidth your entire network, but there's so much traffic North-South and East-West that you create these very, very high latency spots in your network. It's like whack-a-mole every time you change something in an architecture, a new problem jumps up. But talk about a specific example. Let's take Nutanix, for example, where you've got, you know, converged infrastructure, hyper-converged infrastructure. Isn't the network capable within that system of handling this problem? Well, where does it break down? Explain that. So whether it be Nutanix or V-SAN or other things, they, or even for that matter, just any class it could do, they all have the same challenge, which is, if everything is kept within a rack, you wouldn't have a problem. As you go to the second rack, you probably don't have a problem because there's not much East-West. As you go to your third rack and beyond is your chief scale. As you want to stretch it elastically, geographically, then you run into these challenges. And so you won't see these challenges until you go beyond the first rack or third rack, and then you begin to scale. And suddenly you realize you just can't throw bandwidth and buffers at these problems. What you need is a dynamic network. This comes back to the original premise, which is that the hierarchical infrastructure from platform two, client server, was very static. And what you need, and you could build that infrastructure and then pour data and application into it. In this new world, the data sets are so large, they're growing so quickly. The applications grow so quickly. They all scale horizontally. And what you can't just build the infrastructure and pour data and applications in. You need the infrastructure to be built around you. You hear people talking about bringing compute to the data. That's another way of saying, I need to shape the infrastructure around the data. And so the idea that you'd build a static network that's hierarchical and rigid is so 20 years ago. Because the applications and the data sets are so large and growing that the infrastructure needs to be dynamic around that. And that's what a lot of the web scale guys learned a long time ago. Because you're not moving the data. Yeah, you're not moving the data. And these applications are growing horizontally so quickly that you can't pre-engineer the network anymore. The day you cable a network, a hierarchical network, you're banned with latency and everything about your network is determined. And that does not support this idea that you need to build a private cloud that's as agile and flexible as a public alternative. But you're going for scale, obviously. It's not a single rack or a single box. Wikibon's not buying this thing. Not this year. But we're working on that. We have a thing called Switch Zero, which we'll talk to you another day, which is the Switchless switch. Okay, all right, good, we'll get there. But so when you think about the TAM, how do you size this opportunity? It's very large. It's basically data center networking. So it's depending on whose math you use, 30, 40 billion dollar TAM. It's really, if you look at, I mean, pretty clearly Cisco's dominant and Arista's doing a phenomenal job, really being the alternative in that market. And so we're the guy behind Arista by a few years and taking a different architectural approach. Talk about the channel a little bit. We were talking recently about that and how there's big changes. And it's driven by cloud. Obviously open source is changing the whole pricing structure of our industry. But how do you see the channel evolving? You've got the guys who move boxes. That's sort of the historical sweet spot. I presume you're not skating backwards. Talk about your channel strategy, channel partners, your ecosystem that you're building out. What do you see in there? So that's a big part of this year, actually. So it's really great questions. So last summer we signed Arrow as our master distributor. And there's no accident that we did that. And Arrow actually partnered with a lot of the other technology companies that we're working with, whether it be Nutanix or Cloudera or Hortonworks. So there's a new ecosystem of technologies that Arrow wants to bring to market to enable these cloud builders. So the cloud builder customer isn't consuming just a Cisco switch or an EMC array or a NetApp box. They don't buy that way. They're not just looking for best-of-breed product. And the partners are now reflecting that. So there's this move away from the classic box pusher. I'm an EMC var, I'm a Cisco var. To us, they would look more like a system integrator because they think more about the application in the use case and less about the physical infrastructure. They would approach it more from the top down. And so they're trying to solve a business problem and use the right technology to solve the problem. And so they're less about, you know, I'm a Cisco CCIE. As I am, I need to scale a VMware environment from 20,000 to 40,000 to 100,000 VMs. And how do I do that? How do I have an elastic VMware environment that I can stretch geographically and reduce my operating expense? They think, you know, data center in as opposed to component up. It's very, very different. So this cloud builder concept that you guys have been talking about is interesting. So what are the attributes of that persona? Is it an architect? Is it an infrastructure person? Is it an apps person? Yeah, so there, we've named them and we actually have a list of attributes, which I won't go into all of them. They don't give you all of our secrets. But they think really top down. Probably the single biggest thing is they think top down about the infrastructure. That they think about flexibility and agility. They think I'm an architect. I'm a data center architect. I'm a cloud architect. They go by several names that are all related. We just call them all kind of cloud builders. They can be an enlightened networking guy that kind of gets the idea that this hierarchical static network is not really meaning his needs. That, you know, that the scalability horizontally of his user's requirements, if they're, you know, Hortonworks or Clodera, that that network topology is not optimized for that east-west traffic. That there's, you know, these network architects that aren't just skilled in kind of understanding Cisco CLI commands and protocols, but understand the real nature of what a network represents. And what a network represents is an ability to enable applications to do what they need to do, right? Which means I need to meet my SLA. I need to have low latency. I need to, I need to have extremely high throughput. I need to be able to meet these thresholds of SLAs or the applications will then start reacting to the lack of the infrastructure being able to meet their SLA, which ends up potentially causing pathological problems because the applications then start moving applications and data around to work around the fact the infrastructure is behind them. And that can get pathological. So what if you could tell the infrastructure to reshape itself? That's a very, very powerful notion. And because behind all this is the business impact. It's right. Right, which is high there. You know, you probably have enough bandwidth in your network. The problem is it's stuck in a cable over there. You know, so this is part of the problem with the networking guy. I have enough bandwidth in my network. I have, you know, a terabit network. Well, that's great, but it's sort of, you know, the problem is where I want to go, it's congested. It's no different than saying if I'm going between here and New York City, if I take the wrong route, I'm going to hit traffic. And then everybody else tells you, well, I didn't go on the mass bike. I took 95 and they didn't hit any traffic. Well, they were on the wrong road. Tough, right? What if you could know that? That's really the essence of what we do. We can move the bandwidth around to solve your business problem. You probably bought enough bandwidth. It needs to be flexible. The ways of networking. The ways of networking. We talked about this last time. All right, so talk about the market a little bit. I mean, you just did a raise. It's not easy to raise money these days. You're seeing that, you know, you go out to Silicon Valley a lot, I do as well, VCs are tightening up. B rounds are really, really difficult right now. You got to be pleased. You got a, you know, big investor, big name investor. Talk about, you know, what you're seeing in the space, how you guys are doing and just in terms of, you know, funding, what's next for you. Sure. So, I mean, I think we may have talked about this the last time. I'm always raising money. So it's just the nature of what we do. CEO's. That's what we do. So whether it be, you know, this quarter, next quarter, I actually talked to an investor today. So the good news is a lot of people following us now, which is great. So we will continue, we will do another raise this year. And you should raise when you don't need it and we don't need it, so we'll raise. And that's pretty simple. The market for capital is tight, tighter, right? What they want to see is results, right? There's been a lot of stories about SDM and other things. So I think we're past that. The trough of this illusion is almost over. And some of us have actually figured out, actually had to monetize these things. And so those guys that do that and solve real problems will be successful and raise. The challenge now is to really perfect the go-to markets and that's where we're at the focus, so. What's your head count now? We're still under 100, we didn't really disclose that, but we had like 90 people, not a lot. We never really grew too fast. We were ramping sales, but you know, until we see opportunities in different markets, we don't hire, right? So we're very judicious about that. And so we, you know, when I first joined, we had offices in several different places. We shot them all down because it just wasn't prudent. And don't regret any of that. I wish we had presences in some of these places now, but, you know, I can always rent a conference room for meetings, so. So you got a beehive going into Hampshire? Yep, right. It's substantial. It's substantial. And then we have some guys on the West Coast or our partner engineering group, which does a lot of these integrations that we're talking about, whether it be Hortonworks or Nutanix or VMware or NextUp is actually open. And you've done some stuff with Cloudera as well. You've done some stuff with Cloudera, actually, yeah. So having that point of presence out there from the development side is really important to us, because that's where a lot of the technology point is. Okay, so it's kind of a bi-coastal, you know, ratchet game here, is that what we'd expect? Yeah, the right skill out there, you know, we'd love to expand closer into Cambridge, but there's no plans to do that right now, given the current climate. Well, but I mean, you know, there's a lot of uncertainty right now and you guys have, you know, pretty strong vision. So you must be able to attract, you know, certain people. I know New Hampshire maybe is a little tougher than Palo Alto, but Palo Alto is so competitive, you know, with regard to salaries and so forth. So it's a balance, right? No, I mean, especially the skill sets that we go after, I mean, especially now that, you know, we're several years in and there's actually some very interesting projects coming up as parts of our system that we're actually rewriting now. And since we have such strong talent and the architecture is very well-defined, it's easier to get developers now to kind of work on a piece, because it's like, oh, I can own this piece. So it's not hard to attract talent. And so we've been pretty fortunate. And it's what the talent you're looking for is what, systems expertise? It's really software, you know, the prototypical background is kind of operating systems, but they don't really almost teach that in school anymore. Right, so think about connect this API to this API, those are the guys we want. And languages don't matter. Any engineer worth is solved. Can, you know, sadly, I don't speak Italian, but I did program in C, Aida, Bliss, Coball, PL-1, 68,086, 6502, Vax, Myth, Alpha. I know a lot of languages. So languages are not a problem for any real engineer, right? So we're not hung up on that. What we're hung up on is guys that really want to change the world. I believe in the vision and that will just crank, right? Learn from more seasoned guys. So we've had no problem attracting that talent. 6502. Rockwell. 6502. That's good stuff. Yeah, so okay, so you're gonna raise money this year. Other things that we should be looking for, selectively targeting, you know, sales growth, right? Selective target in sales growth, kind of leading the sales charges we expand. We were engaged in a lot of the OEM conversations around, you know, how do we deliver sales in integrated offers. So there's a bunch of that going on. We have some very interesting data center partnerships going on a company called Align Energy that's building kind of next generation data centers, highly elastic data centers, extremely power efficient. And their model is that they have some very interesting power and cooling technology that allows them to have what's called linear PUE. So very efficient data centers as they fill. And so they're very disruptive in the business model because you only pay for what you use in the data center. You don't buy so many square feet and pay for it over four years. You pay for exactly what you use and they can do that because they have key enabling technology. And they looked around for a networking partner and they need a network that's extremely elastic where they can control the traffic in the data center. Who does that? That's kind of what we were born to do. So they program our network to deliver to deliver bandwidth in their data center dynamically. It's a kind of partnership made in heaven and really good guys and they're growing pretty quickly. So you'll see announcements around those kinds of things as we roll forward. And so we're, so data center guys, some service providers more globally, but the big focus for us is financial services and government. Any events that we can find yet this year? We're not doing a lot of events. We're doing local things. So Arrow's our partner. So we're doing a lot of regional stuff where our partners bring their customers and we go to lunch and learned, very focused tactically on the street stuff. We're not doing a lot of events. We'll attend events generally and have meetings around them, but we generally don't do a lot of events. It's interesting. So you're finding much more productivity out of these sort of surgical strikes. With Arrow, recruit the partner, have an event, bring their customers in. We have a couple, you know, have worked with Arrow to do that, you know, bring it against speaker, blah, blah, blah. Have an event, get focused time with them and then, you know, prospect to be very focused. Well, Rich, I'm not surprised that you're on the track that we expected. It's only been 18 months, but you can see you guys are really setting into a group swing here. So congratulations and we'll be watching. Looking for more great things from Plexi. Thanks for coming on. Thanks, always a pleasure. All right, thanks for watching everybody. This is Cube Conversations. This is Dave Vellante. We'll see you next time.