 Live from the Austin Convention Center in Austin, Texas, it's theCUBE at Dell World 2014. Here are your hosts, Dave Vellante and Stu Miniman. Hi everybody, we're back at Dell World. Matt Eastwood is here, he's a group vice president at IDC, focused on a lot of stuff, compute, workloads, cloud, third platform, which is IDC's branding of this sort of next generation of computing. Matt, welcome to theCUBE, great to have you. Thanks, Dave, appreciate it, happy to be here. Dell is a private company, it's all the buzz. You've seen the stuff that Michael wrote leading up to this and the articles and the trade press, he was on CNBC this morning, talks about the great icon and sort of tongue in cheek. From your standpoint, what's it mean for Dell to be private? Obviously a huge news item, but does it change anything? I think the biggest change for Dell is they have flattened out the organization somewhat and they're able to go faster. And I think the big thing I see technologically from Dell since they went private is an enhanced focus on partnerships and kind of deeper engineering partnerships. You can see they're done with Oracle, they're done with Red Hat, what they're doing with Dutonix even. So there's been more of that, but really what the advantage to Dell, or the benefit of being private is really being able to kind of hide some of the impact of the experimentation in the business from the market. So a lot of the re-engineering of Dell that has to happen on the inside really is on the sales side. So can, if you start to think about the market and these market forces around say third platform, mobile, social, cloud, big data, how those come together and how we start to think about selling somewhat different solutions to somewhat different people and also in a different way then that takes different skills. And so there's a lot of focus on retraining the sales force and the go to market and how they think about channel, all of that. And I think there's experimentation happening and you can hide a lot of that under the covers as a private company. There's some, so you've spent a lot of time in the world of compute and generally in servers specifically, I've spent a lot of time in storage. Things like compute and database got kind of boring for a while and they're really getting interesting again. The, I guess the move toward virtualization sort of exposed some of where the value is in the server business and you saw some really interesting times, real focus on market share. Right, that's right. Which I'm sure made your life very interesting. And now the ODMs are coming in to the play and cloud is coming in with what you guys call third platform. What's happening in that world of compute? Well, you have, I mean, so just the very essence of compute starts to change. So the question starts to become, what is a server? And when you start to think about even the roadmaps that are coming at us from say Intel around kind of rack scale and this disaggregated notion of a system getting beyond a physical piece of sheet metal defining a system into something that's defined more at the resource level across a rack becomes pretty interesting. But I think today if I think about the server market and the compute market and what's changing the most is there's really a heavy emphasis on consolidation and virtualization in traditional enterprise data center, traditional workloads. And then of course there's significant amount of growth and capability and capacity happening in the world of hyper scale or web scale for these new generation of workloads, these highly dynamic workloads. And so you start to see where the compute lives and what type of data center environment it lives in and changing pretty dramatically. And that has architectural impact. Do those two worlds collide in your view or do they stay largely separate? Today they're very much still separate. I mean, if you talk to a customer the conversation is still fairly fundamental. It's does this workload fit best in a dedicated environment or a shared environment? And does it fit best on premise or off premise? That tends to be how they think of it. And that leads you to decisions around what the system might look like architecturally. Over time we think that the whole notion of what a cloud is and what a cloud looks like begins to change and it becomes much more seamless and somewhat less visible to somebody that's actually operating a data center. You're going to make a fund account, isn't it? Yeah, yeah, yeah. Yeah, Matt, do you break down by application type? How much compute there is? I guess I heard at a recent analyst segment they said there's going to be certain applications that you're just going to leave in place until you can finally retire. There's some that we're going to migrate over and virtualization helps there and there's some that we're going to rewrite. A lot of attention's been placed to the ODM information that you've put out there. What are some of the numbers and the sizes today that you can break down for? So we do, for 15 years or so, have done annual workloads research where we look at what's actually consuming the compute capacity as well as the storage capacity in the systems. And those broad categories are everything from business processing where you'd have your ERP and your LTP and your Batch to a decision support framework around data warehousing, data analytics, collaboration, messaging, and things like SharePoint and Groupware. You get into core infrastructure, file serving, networking, security, web infrastructures, technical computing. You run that and you actually see that the compute landscape is actually pretty diverse when you start to look at workloads. And what's changing is these workloads are going from say something in the traditional world using IDC terminology, say around second platform that tend to fit inside a socket. So those are workloads that are being consolidated today into these virtual environments. And then you have these more dynamic scale out workloads that are emerging, say the world of Hadoop and of analytics that behave very, very differently and consume resource very differently and actually need to be able to take on resource and give resource back to a pool. So you get into these kind of pool like thinking around the resource. And the third platform, can you say just from a unit standpoint, what percentage of say compute is that today? So I'll take the question this way and say that one of the things we did with our workloads research this year is we made the distinction that you have really too big encampments. You have traditional enterprise and even service providers, the hosters. And you can get to that type of work through surveys and conversations with traditional customers. But then you have these web scalars that look very, very different and you really can't get to them through surveys. And we estimated that in 2013, about just under 25% of all the servers went into web scale environments. Those are the Googles and the Amazons, the Microsofts, the bats in China. And those workloads today are very infrastructure heavy. So what you tend to see from a workloads perspective is to use kind of a traditional analyst tool in a two by two matrix is an assessment of business value versus cost. And there are clearly workloads that are going to continue to fit in that traditional IT model because they don't consume a lot of resource and you can just leave them alone. But there are workloads that are very, very resource intensive and very expensive to manage and maintain. And those tend to fit more in that kind of dynamic cloud like model. So you'll see a shift in the hyperscale of world away from workloads that are really being defined by mobile and consumer into more enterprise class workloads over time. That 25% is, we're talking units, we're talking dollars, we're talking MIPS. MIPS, nobody uses that anymore. It's a unit measurement. And actually the revenue behind it would be less. It would actually probably be more like 15% of revenue because if you think about where the high end systems are, the mainframes in the Unix, what's left of that is all in traditional enterprise. And so today, the server market in the first half of this year has been about 80% x86. And this needs a revenue and 20% non x86, a mainframe in Unix. So still you have a market that's in unit terms overwhelming the x86, but there's a lot of value out there that's in these other platforms. So thinking about the TAM, whatever the IT TAM, let's say it's a trillion, two trillion, some of a huge number. How big is the compute sort of portion of that? And I know it's changing because you get the converge piece coming in but we're talking about what, $100 billion market? Compute market, the way we define servers today, it's about $50, $52 billion. If you add all the pieces together, if you take compute external storage which wouldn't be in the server numbers and traditional networking, it ends up adds up to about $100 billion. About $100 billion, a little bit north of that. And then is it growing? It's growing at a relatively slow rate. So it's called flat. Low, single digits, flat. It's called flat. But there are clearly shifts inside that profit pool. And that shift goes to cloud slash hyperscale and other certain segments within the enterprise? And even maybe to take a different look at it, if you think about what's happening in the market, mentioned ODM earlier, you have very different margin profiles in this market. So if you think about the profit pool perspective, you've got main frames in Unix which are going to still have pretty nice profit margins associated with them. The X86 market has, to the server vendor, typically thrown off 20 to 25 points of gross margin. But in the world of ODM or even the world of a hyperscaler, they're buying systems in the single digit margin range. So it's very different. They don't buy software, they don't buy services, they don't buy financing, so it's a really difficult market to make money in. Okay, so now let's bring that back to Dell. Where's Dell fit? What's Dell's, talk about Dell's strategy. They can certainly thrive on 20% margin. They're happy with that, I would presume. Anything more than that. Seven years ago they went out and created a data center solutions business to go after and to learn from what was happening in the world of hyperscalers. So they started to engage at an engineering level with these hyperscalers. And what they're doing now is taking a lot of the learnings from those conversations and trying to apply it to a more commercial design. Since you could even see that today with FX2, they took some of the learnings from what they've been doing with hyperscalers in terms of how they think of modularity, how they think of density, how they think about energy efficiency and apply it to some of their designs. So for Dell, Dell's always been exceptionally strong in the mid-market. And when I followed them as a public company, you always saw that their margins for Dell as an overall entity were much, much better in when I say mid-market, I mean large, large small businesses to small enterprises. So it's maybe 500 to 5,000 employee-type companies. They also tended to do pretty well in government, but they've tended to have lower margins in enterprise because they haven't had the level of attach there that they get in mid-market. And they've also had lower margins in small s, in small business. So there for Dell, it's about pushing that out and trying to take that footprint that they have with PowerEdge and attach more networking, more storage, more security, more software and services to that. And I presume that's, you would agree, that's the right strategy. What would you tell Michael Dell? Say, hey Matt, what should I do with my core server business and my networking slash converge business? What would you tell him? I mean Dell's in an interesting position because if you think about, to use another kind of overused term these days, software-defined infrastructure, really, and Michael talks a lot about this, the notion of the server in his mind, that's PowerEdge, eating the storage or the server eating the network. That's really all through software. And he's in a position where, to your point earlier Dave, where he can operate happily at a lower margin than maybe the competition that he's going after with that strategy. So if you're a Cisco or an EMC and you're selling software, but also software that's traveling into market with systems that have 50 or 60 points of margin, that's really the target for Dell. Now, the question is really how quickly does the traditional siloed buying in an enterprise change to something that's a bit more horizontal? And generally speaking, I think this market moves in evolutionary terms. So it doesn't happen as fast as people tend to say. But do you buy that notion that the server's kind of eating in? I mean, for years we saw the storage sort of suck value out of the survey. It does feel like it's coming back with Flash and Cisco, what they did with UCS and now the whole converge trend. I mean, it seems like servers is a good place to be. It is a good place to be. And even the term commodity around servers is an interesting one because it's certainly not commodity, if you're Intel and you're getting 60 points of margin out of hands. Hargrove has been commoditizing for 30 years, but you're still money. But I would say that all the players in this market have to have a compute strategy because it is a pretty important component of how this market behaves dynamically. You may not want to have a high volume compute strategy. You might want to have a very targeted, high value strategy. And certainly when you get into conversations around software defined or even converged infrastructures or integrated infrastructures, that becomes a pretty important component. I know you're not a storage guy, but you think EMC needs a server strategy? Or does it already have one with VMware? I think that EMC very much, they shouldn't shy away from playing in computer they need to. And I wouldn't be surprised to see them doing more of that. You know, as they bring VC into EMC and VC, in my view, becomes their converged brand, if you will. I think you'll start to see, they'll have to have some computer, even with Eva. Yeah, Matt, we've talked about from the Viper standpoint to EMC as ECS with the EVO that they have their own server. So in those EMC, not only owns the EMC part, but they own the VMware part. So the question I have for you is, as we look, Dell's placing a lot of bets when you talk about converged and software defined. When you look at things like margin, can Dell make enough money from all of these bets that they're placing there? Partnering with VMware, they've got the Nutanix OEM, they've got Money and Accenta, and on and on from a cloud standpoint, I'm getting questions on Twitter saying, this cloud broker thing, I don't know how Dell makes any money on buying VMs on Google Cloud or Amazon Cloud or things like that. So what's your take on how Dell makes money? I mean, maybe to answer that question, I'll go back to the question that David asked about my advice to Michael. I would tend to say that Dell has gone exceptionally fast as a private company to do a lot of partner-type deals faster than ever, but maybe that gets you questioning about focus and whether or not you're in too many camps and so I think that's a legitimate question and I do think that the way they're thinking about it from an ecosystem perspective in that you have a VMware ecosystem, you've got a Microsoft ecosystem and you've got this kind of open source, open stack ecosystem that they're going after with Red Hat are all legitimate places to be, but your question, Stu, is completely fair because how much opportunity is there for Dell to add value there and capture additional profit pool out of those partnerships is a good question and I think it's early days, it's time will tell if they're able to get enough unique value for Dell to be successful in that type of business. I think this is one of the reasons why Dell had to go private because it's throwing off so much cash from its PC business and it throws off cash from its server business, it can use that cash to make acquisitions in software, it's make acquisitions in services, it's probably got close to a $10 billion services business with the old Perot business and the software business is approaching $2 billion and if they can keep that growing and make more acquisitions, they can shift the margin mix and live off their cash flow. And I think the privatization effort was timed almost perfectly and I don't know that it was completely intentional but the valuation that Dell was able to place on the business when they took it private happened at a time when the PC market was really in the dumps and on top of that they got great value on the debt so they got a great rate on the debt and they've seen, you're right, they've seen the PC market bounce back and I don't know that everyone would have called it exactly that way but there's a great opportunity in the next 18 to 24 months to capitalize on a refresh that's happening in commercial PCs, think about the end of support for XP, even on the server side, even what's going on with Windows Server 2003 and the shift into the need to refresh and update a lot of those architectures. I mean, you're right, the timing of that was perfect because it was timed with the headlines, the headlines couldn't have been worse at the time and so they got the buyer, self Dell, got a great deal in my view and it's going to make a lot of money off of this as is Silver Lake. Silver Lake just has a knack of doing good deals. And they're able to, so far, pay down the debt faster than they actually had budgeted to do and I think they're saying two and a half billion down already in the first half of the year. So the debt will go away relatively quickly in the next couple of years. And they don't have to pay dividends, they don't have to do stock buyback so they shouldn't be constricted from acquisitions. There should be a lot of repatch there to be able to play with and I think that's a, if I think about maybe an area that Dell probably will be very active in M&A and it won't be, I don't see big M&A but tuck in deals with firing IP that they need. So if you think about the way Dell works today, they have a venture's business, they also have a labs business that they've set up recently and they have, so when they think about investment they're thinking about where the gaps are on the portfolio and where they have to tuck in technology. I see security as being an area where there's potential for investment from Dell and then anything data-oriented, data management or data analytic-oriented it seems to make really good sense for them. The security data, the whole IoT things, they could connect into that. I mean they can definitely make a lot of little, they've been pretty successful with their acquisitions. It wasn't, their acquisitions weren't able to transform the company fast enough for the street but you think about the transformation that Dell has engineered for the company. It's pretty astounding and the other thing I wanted to comment on this is they can craft any story they want as a private company. They can mix terabytes and units and dollars and global countries and say, we're number one here and there and really don't have a way to unpack that. You know, I mean that's your job, I guess. But it's a challenge, it's valid. So my advice to Dell has been that they can use and you tweeted on this earlier they can use IDC data and Gartner data to tell a story but I do think that Dell has to face the market maybe twice a year and provide some, what of an update on their business. For the press and for the analysts and for our customers and partners. Right, we'd like to see that obviously. Maybe the European standard with a little less granularity would go along. Yeah, it would be what they showed previously but give people a good peek under the covers at what's going on in the business. Do you think they will? I could see them looking at some of that next year. Selectively maybe. Yeah, they already, I just met with Tom, the CFO and they're already doing this with the debt investors. Yeah, sure. And they still have quarterly calls there but that's more detail than you probably would see if someone like that. So Matt, just a detailed question for you is that Dell seems to have done really well over the last year from the server business but you know, IBM just went into transition to settle in Ovo, the ODM seemed to be coming on strong super micro, you know, underneath. You know, is Dell's server business in trouble? Are they placing solutions in enough of the marketplace and have enough differentiation to keep this growth going? They still definitely still has opportunities to grow and compute. If you look at the server market broadly and you kind of ask why IBM picked now to do what they did I would actually argue that IBM should have done this maybe two years sooner but it's a couple things. One is the overall market for X86 servers is more predictable now than it was five years ago. It's flattened out, right? And it's eaten into the market, it's at 80%, is it going to get to 90, sure. But it's going to take some time to get there as you eat into those high value workloads. But really, you have a couple things that have been happening in the server market that have been interesting. One is Cisco coming into the market and now having about five points of share in the market all up but putting pressure on the higher value of X86 on IBM, on HP. Then you have the emergence of China as a market which is really sourcing from local Chinese OEMs from Huawei, from Lenovo, from Innsbruh and Sugen. And you have the emergence of the ODMs as well. So you have these kind of three forces that are working on the market that are putting pressure on the traditional competitive ecosystem. Well, and right now Dell's capitalizing on the turmoil IBM X86 sale. I mean, the resellers were not happy about that and we all know that story. But Lenovo could emerge as a really formidable competitor here, I think. I think they can and I think they will. I actually think what will happen is there will be a period of transition where IBM, the part of the X86 market that IBM really always aspired to have was the higher end. Scalable X86 blades. And that's generally where they carved out their niche. Lenovo I think will take a very different approach. They're going to very much focus still on convergence but they're also going to go very hard at volume and build a volume business. And that volume business will also include hyperscalers. So I think you'll see a very different Lenovo emerge in the market over the next few years. And of course there'll be some places where they lose out. I think there'll be some nationalism, some government type buying, maybe HPC type buying that won't stay with Lenovo that will go to competitors. But I think the upside for Lenovo is more than the downside. And you will see some adjustments in the portfolio over the next two years. But if you look out three to five years, I think you'll see Lenovo as a very formal partner. And they align more interestingly with Dell than say they did in IBM or HP. Well, anyway, overall great assessment and session here on sort of unpacking the business and looking at Dell's prospects, Matt. So we've got to leave it there. I really appreciate you coming on. Thanks for having me there. Appreciate it. All right, keep right there everybody. We'll be back with our next guest. This is theCUBE. We're live from Dell World 2014. We'll be right back.