 Okay, we're back. This is Dave Vellante. We're live at EMC World, Silicon Angles Continuous Coverage, wall-to-wall coverage of EMC World. We're here with Mike Capellis, who's the CEO of VCE. Mike, welcome back to theCUBE. I don't know if you remember, but our very first CUBE was two years ago in Boston at EMC World, and you came on with Joe Tucci. We had a great conversation. My co-host asked Joe if storage was sexy and Joe didn't really think so, but. So welcome back. You know, I actually remember that. We were telling the story about how I ended up as I was mining my own business and then Joe into a baseball game and somewhere between one beer and one hot dog, I ended up, you know, working with him at VCE. Yeah, so we're also here with Stu Miniman, who follows converged infrastructure for us. So Michael, I was just in your session upstairs with the analysts. I thought it was good. I got some good questions out of that. I mean, analysts sometimes ask good questions, sometimes not, but I thought your brief history lesson was good. You used to talk about VCE, it came out in 2009, and what I always liked about listening to you talk about this is you have a practitioner's perspective. You know, you're not just a marketing CEO talking points. Talk a little bit about, take us back to the vision that you guys had with VCE and why that's so important from a practitioner's perspective. Yeah, so I mean, so VCE was actually formed, those of you who don't know, it is it was a venture that was formed between EMC Cisco with participation by Intel and VMware, and people ask how it sort of originated. Well, it actually goes back a long time, and I have two parallels that you make. If you think about the world of networking, you know, we're around networking and we had fixed networks and data networks and point to point and those see whatever. And you know, the IP networks came along, you said, I'm going to write one place and I don't have to know or care where the physical layer is underneath it. And that spawned, oh, probably 10 years of extremely rapid technological advancement. Well, between both companies, between both Cisco and EMC, we said, well, we know that there will be a convergence of where the network starts and compute ends. What's the blade? What's the switch? How storage fits in? And then of course, EMC through Joe, getting together with VMware and how virtualization. So when you looked at the pieces, you had compute, networking, the interconnect, storage, virtualization, the connect. And we said, we can develop an alternate model, which is very quick, not constrained by the things normal big companies and be able to create really the next, literally x86 class mainframe by using the best of capabilities. And everybody's familiar with big mergers. I might know a little bit about that. And so it was an alternate way to do fast innovation using the best technology components under the understanding, just like networks, there would be convergence. Yeah, so I didn't realize, I guess, I thought about what you just said about x86 mainframe. You guys actually copied IBM on that front, not with the x86 mainframe, but the mainframe, right? So in a way, you were trying to conceive of this software mainframe, as Maritz calls it, with the x86 base mainframe. Now I wanted to follow up on that. So 2009, you announced VBoc, IBM came out, I guess it was last month with three years later, with their converge system, integrated systems, pure systems. So you guys, that's a three year time to market advantage that you have. Now there are certain things about IBM's system that I like, codifying all that knowledge. Now I'm trying to figure out how much of that is real, how much of that is good marketing. Can the industry actually take those learnings, codify it, and embed it into systems so that those systems are intelligent in a way that can align to applications and workloads? Yeah, I think you've actually hit on the two most important things. Everybody talks about the different offerings in the marketplace, you know, everybody's xconverged, somebody's reference architecture. What really defines this is you really do have to look at applications and workloads. The answer is, well, what do you use? Well, it depends on what you're trying to do. So if you take what IBM has done, and by the way, I'm incredibly respectful of them. They've been around a long time, and obviously their reputation speaks for themselves. But they were trying to say, is we have an understanding of the software stack that allows you to be able to take and to deploy applications in a very robust way, ensuring all the things you need from mainframe class, you know, reliability, robustness. They then took, and they put their cloudburst layer on the top of the stage. Now we'll add some feature function here to give it the kind of things you need for services in a cloud layer. The ability to do dynamic portals to be able to do monitoring and metering all that stuff. So you know, I mean, one thing I like about what IBM does is basically when IBM jumps in the fray, they have validated the market and high tide floats all ships. So this does say they will approach it from their stack, which their knowledge of them, they certainly understand the applications world, will approach it from a view of saying, we have taken the absolute best in class components. You know, Cisco with their innovation and lead and networking down through the fabric and into the Intel blades, storage, virtualization, and we will approach them both sides saying, this is a new way to deploy infrastructure. As you need it, where you don't have to know where care with that physical areas, then let the applications come and determine what the workload is, depending on what the need is. Yeah, so the validation point is key. And now when EMC quietly put in its 10K, the numbers that they had to report because it was getting so measurable, Stu and I sort of squinted through those and we tried to reverse engineer them and build a little VCE model. For the purpose of saying, well, how big is this thing really? And it's quite large. Now, of course, the other point is, you guys are investing a lot of money into it. And so, if you looked at it from a VC standpoint, wow, this is really an expensive endeavor. And so what we did is we said, well, let's look at the market. Why are people doing this? The market's absolutely enormous. We quantified a $400 billion TAM because it's everything. It's compute, it's storage, it's networking and it's the services and software around it. And we made a point that the services and concepts should decline. But still, we're talking about an enormous market and it's almost like government's investing in things. A little tiny startup can't do this, you know? Well, so first place, this gives me an opportunity to set the record straight on one thing. And God knows the audience doesn't want an accounting class, but I'm going to offer one anyway. The way the business model works is, when we sell a V block, and this is fully disclosed in all financials, is EMC picks up the revenue on the margin of their products. Cisco picks up the revenue margin of their products and what is left is reported is only the expense. And as Joe Tucci says, the bigger the expense, get the more successful we are. So people sometimes look and say, well, that is actually the profit. It's not, the profit's already been taken. And internally, VCE has like any normal company, a P&L based on a stand-alone. So don't confuse the reported expense number from the profitability. And we didn't, that's what I say. We tried to do both of them. Now, back to the relevant point. Because that's an important point, because what you're saying is that Brian Gallagher's P&L takes profit from that and Rich DePolitano's P&L takes profit. We had Rich on earlier. He's saying, wow, I love V block. I'm making, it's a big chunk of my business. So that's, we get that. And that's a good thing. All right, so anyway, continue. I'm sorry. So the point you were making is how big is the market? Well, the question you want to ask is, at what point and at what adoption level do you have where converged infrastructure starts to represent a relative percentage of the market? And so what we're already starting to see is you start to see where somebody will make a decision to actually deploy applications this way. And that decision will actually sort of define what they do in switching for the entire data center, server for the entire center and compute. So I mean, we look at the compounded growth rates. One of the reasons, I'm kind of a common sense guy, is why is that the case? Well, the value proposition says, I can deploy it. I don't have to do any release management. Don't have to do any configuration management. I can actually start to think about just deploying infrastructure, which is really the cloud model, when I need it, as I need it. The V block has been optimized so that it is in perfect balance relevant to most application loads. So when you try to do it on your own, you're always over by the storage, under by the compute, and you never get it. So the reason why it continued to grow is customers are saying, the goal here is the fastest possible deployment of applications, the most flexibility without having to know where care with the infrastructures, and it's a compelling common sense value problem. So, go ahead. So we talked about the different components that make up the converged infrastructure, where there's really been an opportunity in the market is from the systems management, rather than individual product management, and the convergent should change the operating models. So can you talk a little bit as to what VCE sees as their role there and where are you driving things? Yeah, so what is really interesting is you have to think about things differently when you start assuming virtualization is the way you deploy. In a traditional style world in systems management, it said, I have an application and I have a system, and the way Dena said to operators, I have clusters. I have a mainframe cluster over here that runs some financial applications. I got a Windows world over here that runs some web serving. I've got a Unix world over here that does something. And so we built system management to say we have a one-to-one relationship between the application and the system. Throw 1,000 VMs in, it doesn't know anymore. And so we have actually taken a very different look at systems management by saying I have, at the element level, or defined as the compute engine, the network level, and the storage level, I have all the information about the components. Let me just extract and identify those components. I'll pass that information up to the virtualization engine which says, now what do I need? I need to know which resources deploy, I need to know how to do it, I need to be able to address what pools I want to use. So we took an opposite world, instead of looking down here going up, we said let's look at system management from what does the virtualization layer do, go down, automatically discover and literally strip out layers that you don't need anymore. And so the answer to the question is, well, what do you do for systems management? Is we inherently take advantage of all the capability that's already there and only add those things to the virtualization that it needs to do resource management. It's a very different view. I thought Andrew Reichman from Forrester asked a really good question. He didn't close it right, but I can say that. It's easy for me to say, but I thought his question was one of the better ones I heard. So I'm going to re-ask it here and maybe talk about it. He said, you guys talk about best of breed. And he said that, and I buy that EMC storage and Cisco networking, but he said the Cisco compute piece is somewhat speculative. That's the way he said it. And then he asked, this is where he blew it. He said, is that holding you back? And you were very nice. You could have said, we're crushing it. What do you mean is it holding us back? So that was the wrong close. But what I thought was, could it potentially be a barrier down the road? I wonder if you could address that. So the way we answer that question is that, first place is, let's remember the way we think about this, and we do think about it differently. We don't think about it as just a blade or a processor. What the UCS is, is you start with very dense, high performance IP switching. You put a fabric into it that allows you to go down to the compute engine and you actually have a zero latency configuration between network and compute. At that point, we certainly can adapt through Intel and our partnerships. Intel technology adoption, the fastest possible rate. To that, we embed storage and we natural virtualize. The question isn't, just is it, do you have a better faster blade because Intel basically is providing the core? We are actually delivering something very different. It is the integration of the switch with the fabric, with the compute, to deliver an integrated approach which dramatically changes all the value proposition. So, somebody else got a different question. If you only want to run a very low end server as a single application, it's probably not the thing. If you want to manage applications and actually create compute as a utility, then it's a very different question. And the other thing you said is you're not going to do, some of the things that, some of the workloads and applications that V-Blocks is going into, you're just not going to do them on a reference architecture. Well, again, I always like to differentiate the market. A reference architecture, and we're fine with that, Cisco EMC is supporting those, is a very different thing than a system that's integrated. If you want to go out, one of the analogies I mean, would you ever build a mainframe on the shop floor and then try to run mission critical applications? Probably not. Would you take some storage and some switches and a server and build it on the floor and be able to run a Windows, maybe you want to run some exchange or SharePoint. So, there's a set of class applications or even some customers who have their own IP they want to apply, that will be for some class of application. V-Block is actually something very different. It's an integrated system that you can run, mixed workloads, and it is in fact a much higher abstraction. And you know, which one you run where? Well, it depends what your applications are. If you want us to go provide a V-Block simply to run exchange a little faster, you know, you're probably going to buy something else. So, I'm wondering if you could address the service provider market, because we see, you know, rather than, most enterprises have multiple applications, service providers very much a scale out architecture, but I understand you guys have been doing pretty well there, so maybe you can address how that fits together. Well, depending on what class of service provider you look at, because you have SIs and so on, but let's just use what would be the traditional. If you're a service provider and you've been running a network, what is it that you used to do? You manage the network, you put IP on top of that, and what you sold your value, your customers, you don't have to know where care where the network is, I'll do billing for you, I'll do quality of service, I'll be able to differentiate SLAs. So the first thing I said, well, if I'm there, could I now start to sell other higher level forms of compute? Drop in a V-Block. V-Block goes in, you can now sell the entire infrastructure as a service, not only with compute networking and storage at that level, but the actual network underneath it. You have your tools for the differentiate quality of service, I got that, I can bill it. So it's a natural, and then you can consume it as you need it, if you need more, you add another V-Block. So it's a natural environment for the service providers, and by the way, as you know, these are very technically bright people who are used to running things at enormous scale. And I want to come back to this, so service providers, you think service providers, you think of large infrastructure and repeatable, so I want to come back to the reference architecture discussion, because I personally think that the reference architecture, you guys doing a reference architecture would dilute a core value proposition, in my opinion. At the same time, the channel's clamoring for reference, and reference architecture is a bigger business because you could define it pretty much any way you want. My question is, are you going to do a baby V-Block, and can you talk about that a little bit? Yeah, so we agree. So the way we sort of think about this, and one of the things we're particularly proud of is, particularly at some stage of your evolution, you want to make sure you have a great customer experience, and that's something we're particularly proud of. So what we said our first foray into that was, we have a bunch of large customers who actually want to do a remote office deployment. So we said, aha. So we have been developing a product, which is a lower end product, which is for remote distribution to be used with our bigger engines. They're called 300, 700, but whatever they're called. That's step one. What we'll learn from that are the characteristics of what you now need to do to build a V-Block, which is stand alone, which has different requirements for manageability and some other things. We have been working with our channel partners to define what are those characteristics. So the natural evolution, by the way, of all technologies. You come in early, you get enterprise customers to validate it, you then start to vote for the channel, you then validate that, you develop a remote office integrated system. From that, you develop a more stand aligned product, you take that product through the channel and offer it as a stand alone. So this is classic evolution, step by step. We will be there. We're not going to pre-announce the product, but we are in active discussions with partners on requirements for a stand alone product in that market. The second reason to be honest with you is we've actually had so much volume and scale. Demand hasn't been a problem so far. So I wonder if you could speak a little bit more to the channel. There were a little bit of bumps in the road, I'd say, kind of the first 12 to 18 months. I agree with that. And kind of the whole Acadia model got sorted out to VCE. I think signals we've gotten there, things are better, but there's a big fight in the channel for who's going to own that piece. Well, so when we started the model, it was a service model. And I mean, I'll have to admit, I kind of had a point of view on that. Having a service model is competition with the channel. So, okay, we're not going to do that. That's where that business is. So I was bumping the road. Second bumping the road is we had a reference architecture for a very complex product. And what is the classic step of channel deployment? You have to mature the model. You have to mature the product. It has to be repeatable. You have to have tools. And so as we came up that curve, I mean, it is classic. A very complex product in early stages is not terribly productive for the channel. A mature product with services value added content with tools is pretty good. So we came up and we got a little bit of positioning by different parties. We're also doing a little bit better of being much more consistent in, this is what we're going to do with not, you noticed how clear I was. This is what we're going to do on the commercial product. It's extremely clear. So we had some bumps in the road. Part of it was just maturation. And I do think right now we are recruiting partners. We're very clear. We're not going to compete by offering services at the end product, you know, for an SP. We're not going to offer software directly. That is for our ISVs. And we are going to offer infrastructure and a deployable model for the channel. But this is a channel model. 2020 hindsight, I think we could have done a better job of communication. We could have stepped it a little slower and we could have matured the processes. But it is a lot better. Yeah. Mike, you seem pretty charged up about this current role. I mean, you're a CEO, a compact, MCI. I mean, but you seem really passionate about this initiative. I mean, you having fun? Well, first place, you've known me a long time. Do I ever not have fun? There are a couple of WorldCom days that weren't quite as much fun as some others. But you know, these things happen. You know, bankruptcy and fraud weren't the greatest technological attributes for a brand, but we kind of got through that. So I have a couple of views on this. What do you want to do? You want to do something that's really, this is really moving markets. It has the great benefit of having Cisco EMC and VMware. Could you ask for better technological partners to scale it? And you do also like to work with people that you like a lot. I've known Joe and John a long time. So you know, it has been relevant to the market. I love working with the kinds of young teams and creating the culture of a can do. And so, yeah, sure it's fun. That's great. And I know you rock and roll aficionado. So we actually, we were down at Sapphire last week. We heard Van Halen. I hadn't seen them since the 80s. And we're okay. So your favorite U2 song? My favorite U2 song, I would say Bloody Sunday. It's my favorite U2 song. Every once in a while it's a beautiful day in the BC world. Yeah, you know. How about your favorite rock song? For me it's too cliche, I don't know. But do you play it? Do I play it? No. No, no, you play? No. Well, I'm a hack. Yeah, okay. Yeah, favorite rock and roll band. Favorite rock and roll band, Clash. All right. Love the Clash. There you go. So we've concluded that. How are you? You know, I like just about everything. And everybody says that. But I mean, some have been about hack on the guitar. The first song I ever learned to play on the guitar was Green Day. So that was the place I kind of started. You know, it was kind of adorable. But I go everywhere. I mean, you know, the next one I think was Life Outs was the next song I learned to play. So I kind of go all over the place. So who's the entertainment this week? Maroon 5 this week. Maroon 5? Oh, okay. Staying last year at Sapphire was very good. So David Lee Roth was the lead singer. But his voice is going. So they had to crank up the guitars and you couldn't really hear it. All right. So we're getting low on time. I do have one more question I want to ask. I'll be back. Ask something beautiful here. Okay, all right. Well, come on. So VCD is up to 1,000 people. How do you maintain that kind of startup atmosphere and keep it going as you kind of, as you said, stabilize the product line? Well, I think what happens is, you know, you have certain things you believe in and you kind of, as you bring people on board, you kind of get a sense for, do those attributes work or those attributes not work? And, you know, while it's 1,000 people and you continue to work on, it is very clear that this is kind of, you know, these are, you know, everybody says, well, how do you move fast? One of the ways you move fast is everybody we hire has a pretty strong technical orientation. So I don't have to sit here and explain what it is. So we have certain things that we look for. And that is that you got to be passionate about this business. Just to come, you have to assume you have some amount of, you know, I'm a bit of a maverick anyway, a bit about, you know, a little, you know, sort of we'll take some risks on and you have to be passionate about making a difference. You know, over time, you know, everybody doesn't have to look exactly the same because culture is about diversity of thought. But you do sort of get a feel. And right now, it's a place people want to come work. All right, Mike Cappell, to me, the embodiment of strategy and execution, congratulations on the progress that you've made. I really appreciate the candid conversation and thanks very much for coming on the queue. It's really a pleasure to have you. Thank you. All right, keep it right there. We'll be right back with SiliconANGLE.tv's continuous coverage, live from EMC World. Be right back.