 At VMworld 2012, this is in San Francisco. This is siliconangle.com, siliconangle.tv's flagship broadcast theCUBE, where we go out to the events and extract the signal from the noise and share that with you. And we've had a great time here at VMworld, day one of three days of coverage, extended continuous coverage. Pat Gelsinger took the mantle today past the torch from Paul Moritz and we're excited to do that. I'm John Furrier, the founder of SiliconANGLE and I'm joining on this segment with my co-host, Stu Miniman, Stu. Hey John, welcome to theCUBE. Great watching you and Dave kick things off here, another VMworld, great vibe and excitement here in the hang space where all the cool people are, you know, the bloggers, you've got video games and air hockey and lots of sharing and networking going on. And excited to have Dheeraj Pandey, who's the CEO of Nutanix. So I first heard of Nutanix a little over a year ago, talked to them a little bit last year at VMworld and they're an exciting company in the convergent space. And John, you've been talking this morning about going beyond convergent infrastructure to data infrastructure and Nutanix is right at the heart of this. So we're going to talk about, you know, infrastructure, big data, VDI, I mean, all the big buzzwords, you know, cloud maybe even. So Dheeraj, welcome to theCUBE. Yeah, thank you Stu and thank you John. So give us a quick update because you guys had just raised another round of funding, Series C. The company's doing extremely well over the past year. I mean, the trends are interesting. They're, you know, the winds at your back. I mean, you got cloud, solid state, big data. Everyone's recognized that this new era of computing, this new modern era, data center, big data, end user applications. So give us a quick update. Sure. Latest round of funding, the details, number of employees, company status, how much you're growing, et cetera, et cetera. Sure. The company was formed in September of 2009 and really conceived of the idea in December of 2009. So it's about two and a half years since we've been doing this. We raised the series A and B in 2010, 2011 in summers of 2010, 2011 for about 38 and a half million. Then we just raised 33 million in Series C about four days ago for a total of $72 million that we've raised overall, out of which we still have 50 plus in the bank. So this was an opportunity for us to go and flip the switch in the business. So we got some really good feedback from the market in the first year of selling. We sold a whole lot more in three quarters than some of the big public companies that you know of today in the first six to eight quarters for them. About 105 people in terms of employees started to expand to Europe and Australia and Japan. And I think this is the beginning of what we believe is a company of lasting value. So obviously some hot, just to let the investors and advisors, you guys have a great team around you, obviously from an institutional investor standpoint, some fantastic investors. But I'd like to ask you specifically to tell, share with us why it's so hot right now. I'm not talking about the frothy VC climate, which it is, but you know, this is a perfect storm right now in this modern era. You have transformation, need and demand. You have technology explosion with solid state and convergent infrastructure, as we're calling data infrastructure and big data. Explain to the people who are in IT perfection, who's not in the trenches, not inside the ropes, if you will, why this is happening and what your view is. Yeah. So I forgot to mention our investors in series C. Our existing investors actually came in with a big round and then add to that was Goldman Sachs and battery ventures. Coming to a question of why things are hot. Well, the biggest thing is the ubiquitous x86 hardware platform. Everything will be x86. So what does it mean for the rest of the infrastructure? Well, everything has to actually go and tie with x86 and virtualization is going to be the substrate that will decouple hardware from the rest of the software. And we're doing this exact same thing for storage as well. We're saying storage controllers cannot and should not run on bare metal. They should be running inside the substrate, which is this secular trend that we're seeing in data centers. You saw that with networking, you're going to see that with storage now. But what does it really do to the end user? For one, obviously it reduces cost because now you have this one common way of actually building data centers and everything is in software. But more than that is the ability to simplify. What we are saying is that when you collapse computing storage together, you are really simplifying data centers. And today, IT is chronically understaffed and they're all looking for ways to simplify the jobs. They want to go back home at five p.m. if they could. And convergence is all about simplifying the data center. Now you get a lot of technology benefits that it would get into. But the big thing is this is a building block and you basically start small and you can grow it as large as you want. And that's it. Stu, you've been covering obviously, at Wikibon, you're the lead analyst on what I would call convergence restriction. Now we're going to be rolling out a series of editorial and research around data infrastructure, which is just different. And we had, I think the poster child for this is really going to be David Flynn from Fusion who just said earlier, if data, the notion of solid state was around during when people designed OSes, we'd have a different look of what an OS is. So I truly believe, I think he's validated that this new environment is going to be the operating environment. You know, abstract, pool, and automated. I mean that's operating system code words right there. So Stu, your angle on this, what is your take? Why is data infrastructure different than converged infrastructure in the sense of it's like, it builds on the shoulders of converged infrastructure? Right. And what's different? So if you listened to the keynote this morning, the discussion we've had at VMworld for the last year or two is, while I'm an infrastructure guy, we need to understand that it's about the applications and you know, where's the real value being created in the market? You know, Dave Vellante always says, if you're a storage administrator and your value is creating and changing lones, you're going to be out of the job in a couple of years, so you need to be able to grow from that. And we've had in convergence, the discussion's been, you know, who built this great performing, you know, cool hardware? And actually to bring it back to Nutanix, one of the things that really impressed me when I first heard them is, you talked about Fusion IO. They use Fusion IO in their box. They use standard SSDs. It's all off-the-shelf pieces. So, dear Audrey, if I can throw it back to you, you know, how do you differentiate in the market and where do you create your value and how are you sustainably differentiated in the market if you're just using, you know, off-the-shelf, you know, x86 components, pieces you're working with, you know, merchant-silicon switches like Arista, you know, how does that work out? The DNA of the company is Google File System, Oracle Exadata, Big Data, AsterData, which was recently acquired by Teradata. We believe that there is a way to embarrassingly parallelize the data center. You know, you have a way of looking at, well, just throw another server. It has enough storage, enough compute, enough memory, enough Fusion IO, and you can have apps that actually grow and scale in this kind of environment. And that's where we actually differentiate from pure play convergence boxes, which are like these massive monoliths which have a big-something storage, a big-something computing, and a big-something network. We're saying you should be able to really do this at a micro level. You know, you should be able to do this at one-fourth of a 2U, not really a whole 42U birth of a rack, but one-fourth of a 2U. But when Nutanix comes in and really provides that special sauce is all software, we're saying we never use hardware as a crutch. So if you have to implement high availability, will you do it in software? If you have to implement rate, it's going to be in software. If you have to implement fault tolerance, it's going to be in software. If you have to implement performance, it's going to use commodity-off-the-shell stuff. So that is the basic premise of the company, that everything that we do in storage will not use any custom hardware facilities to really deliver all that. I wonder if you have a thought on how much lives in your software and how much lives in the virtualization layer. A lot of discussion this week about, you know, how much feature functionality is VMware going to kind of suck in and make standard in the hypervisor? What's your take on that and how do you interact on that piece? So we definitely believe that VMware as a virtualization substrate is very important for compute and memory. I think networking, you already see the piece that Nysera was, for example, building big switch on the SDN vendor. So, you know, VMware went for an acquisition to say, well, we can bolster our virtual switch story. I think storage is a big piece. And up until now, it has been about going and doing like singular IOPS, as opposed to looking at big distributed systems problems. We believe the biggest value will come in distributed systems. How do you really start small and grow incrementally and grow to massively large clusters all using one single system view, one single namespace? And I think that could live in the storage controller, which is not inside the hypervisor and it could still sit on top of the hypervisor. So when you look at applications that go in a virtual environment, things like all the Microsoft exchanges and SQL and then SAP and even Oracle, we've been talking about for a couple of years, the new one that's kind of hot and VMware's jumping in is big data. You guys are already putting big data in your environments. Can you tell us, is that a separate silo? How do you put it together? What big data applications are you doing? Is it Hadoop or just or other stuff? Well, big data's biggest premise was no SQL, no SAM. They were trying to collapse the infrastructure, flat infrastructures, flat data centers. You run computing storage right there where it belongs. You bring data as close to compute as possible, which is exactly Nutanix. The only thing is that we bring the virtue of virtualization, x86 virtualization. And that is what VMware is also talking about with Project Serengeti. They're saying, look, IT, if it needs to manage Hadoop clusters, they need to do this in virtualization substrate as opposed to bare metal. If all of Hadoop in the public cloud is running in virtualization, it should be running in VMs in the private cloud as well. We go in there and say, look, it's great that you actually realize the value of x86 virtualization for Hadoop environments. The next level is, how do you bring all the enterprise-grade goodness that the storage guys have been so good at in the last 15 years? How do you bring all of that into the server? And that's where we come in. If Hadoop has to become mission critical, it needs to have snapshots and disaster recovery and fault tolerance and cloning and backups and fusion IO goodness, all that stuff has to come in to the server. And that is where Nutanix comes in, which makes it an enterprise-grade big data deployment as opposed to what developers are basically kicking the tires off. So I want to ask a question around this whole hardware as a crutch comment, which we want to drill down on because really that has always been the case, right? People bought this big EMC, Hunk and VMAX, and these Sands are all over the place. We were at the Cassandra Summit doing theCUBE, and they're doing a lot of pioneering work with Solid State and with Cassandra, although we're more of a Hadoop shop ourselves. I talked to a lot of guys saying, hey, we're going to put in two huge, multi-million dollar, two SAN orders, and we just basically eliminated one, and that paid for an entire set of projects because all we did was put Solid State in front of some commodity hardware. So can you talk about that? Because this is not just a one-off. I've heard this a few times where, and this points to this modern era of reconfiguration of this new operating environment where the data center of the future is not going to be so much off-prem as much as people thought. It's a lot of on-premise innovation. Can you drill, and this is just one example, can you talk more about that in other use cases that you're seeing? Yeah, so big data could actually, you could think of it in capacity, you could think of it in performance, you could think of it in terms of scale. It's all of that, right? A one big monolithic box that you bought might have been good enough for a pilot, but it might not be good enough for production. Also, you don't want to keep it on your books for like 12, 18 months before you realize the value of that million dollar procurement. So what Fusion IOs the world have actually started to talk about is decentralized data centers, and how do you take that one big monolith and really kind of scatter it among large number of machines, which we believe is the data center of the next generation. You really want to do this in a distributed systems manner, where you don't think of big monolithic boxes, but instead think of fragments, shards, that you buy at $20,000 as opposed to $2 million, when you need to grow, you grow. And it goes back to this customer propensity of saying, well, that's the way I buy my Salesforce.com license. Why should I buy hardware differently? If I pay per month per user when it comes to software-to-service SaaS applications, they should be able to do the same for on-premise hardware as well. And I think that OPEX goodness and that ability to actually start small and grow as you go is what's driving all of this stuff. What's the biggest disruption though from what you're seeing? Because you guys are out forging these new use cases, you're actually in some cutting edge deployments, and then that's just pilots. There's some production stuff going on there. Yeah, so the biggest disruption is being able to take this simplified, flattened data center infrastructure and sell it to the application guys. Because these are the guys who eventually get shouted upon the most. They're like, well, you have a deadline to meet and I don't care what your storage guy's saying, what your networking guy's saying, what your virtualization vendor is saying, I want you to finish something like a VDI project, like a Hadoop project, like a server virtualization project in X number of months. So like the way NetApp did this early on, they took it to the have-nots. They said, here is an appliance, simple story, and you guys can actually break away and do your own infrastructure. This is exactly that story. We are taking this story to the application guys, to the cloud guys, to the virtualization guys, and saying you can build a VDI cloud in half a day. So in the spirit of the keynote, which talked about kind of the old way, and these are my words now, old way, new way, PCs to mobile, old apps to new apps, servers to cloud, talk about the old way and new way. I mean, basically kind of when I hear you talk about that, I think about shadow IT. I mean, I think why shadow IT exists is to get around the slowness of the incumbent, either staff or hardware or whatever, right? So they work in the shadows, but now with accelerated timetables for application deployment, this modern era speaks to this, right? So there's all kinds of side effects, compliance, for example, and many more. We've documented on theCUBE many conversations. Can you talk about that evolution just as an organizational dynamic and how do you sell into it? Is there processes for this or is it new? Yeah, I mean, it's not unusual. I mean, as I said, NetApp went after work groups and departmental guys. VMware initially was selling desktops to workstation users and things like that. Eventually they started doing test and dev. So they went after the low friction, high velocity business. And I think a lot of these CIO driven initiatives, which actually need to be high velocity and project driven, need to actually come that way. Now, what's the biggest value add? I've talked about technology and fast and scale and all that stuff, but the biggest stuff is peace of mind. The app guy picks up the phone and says, look, at three AM in the morning, your PhD should be working on my problem. I do not care for whether it's VMware's problem or it's storage problem or networking problem or whosoever, it's your damn problem. So, Dhiraj, final question I have for you is what does the organization that runs your box look like in the future? I mean, we used to have, as we said, the silos or the cylinders of excellence people sometimes said. Well, I'll tend to look at, there'll be way more generalists and there'll be very few specialists. So it's like healthcare. You have a lot of family practitioners and primary care physicians and they'll actually take a look at most of the stuff because it's simplified, virtualization simplifies a lot of things. And then when they actually debug a problem and say, well, you need to go to an oncologist or you need to go to a radiologist or whosoever is where things will happen. But the specialist's role will actually diminish. The generalist's role will actually really become bigger. My final question is, what does data infrastructure mean to you in the sense that if you could share your perspective on data infrastructure being defined as building on top of convergent infrastructure in kind of an older class and with data being the central value proposition and with these new explosive use cases around mobile, cloud, and social, et cetera, what is your take on data infrastructure? What does that mean to you? So the biggest thing is velocity, right? Because as you said, application deployment times have actually reduced. It's about how quickly can you bring up something and provide SLAs to the end user and it's about performance, reliability, and everything else taken together. So if you're building an infrastructure as an IT organization, you have to figure out multi-tenancy, you have to figure out how you have one shared hardware that doesn't commit to policies early on when you buy it from the vendor, but instead says, look, we provide you mechanisms to do RAID and fault tolerance and HA and FusionIO and all that stuff, but it's up to you how you take and apply policies on top of it. So it means you will differentiate in terms of your expensive customers versus your cheaper customers and say, well, I'll back you up every two days, I'll back you up every two hours and I'll snapshot you every this frequently and I'll give you the FusionIO love and I won't give you that FusionIO love. All that stuff, disaster recovery policies, HA policies, snapshot backup policies, all those things are defined by this infrastructure administrator. And they're saying, I'm not bound and behold into this hardware that only does one and one thing only. That, I believe, and that's what people are calling software defined storage, they're saying, look, you have mechanisms that the vendors provide, but you have policies that the admin actually goes and administers at the time of deployment. Awesome, well, I'll just ask one follow up. This is my truly last question and then we'll take a break and then we've got some time. What is your take on VMware as a company right now? Where are they right now in your mind's eye and where they are in terms of a grade and kind of some color around as a company how they're doing and then secondly, behind that, what do they need to work on? Is it end user computing? Is it filling in the holes of vFabric and vMotion? What's specifically, what's on their to-do list? Well, I think VMware is where Oracle was in the year 2000. The king of the hill. They owned the database like Oracle did in 2000. VMware owns the virtualization space today. But then Oracle got cocky. And also, and Microsoft became good enough for the mid-market, right? So by 2004, SQL Server was everywhere in Windows deployments and Microsoft knew how to bundle it with the rest of the Windows and everything else and that's gonna begin to happen to VMware now. You'll see that Hyper-V is gonna become good enough for a lot of the guys. In fact, the database market kind of tri-focated into the mid-market of SQL Server, the global 2000 of Oracle database stuff and then the web guys who did MySQL. Something similar is happening to virtualization with KVM being the stuff for the service providers. Then there's gonna be Hyper-V for the mid-market and then VMware for the global 2000. I think what VMware needs to work on is focus. I think Oracle is really good at that. They knew that they'll never be a consumer company. Larry Ellison could have bought Apple because he sat on the Apple's board in the year 98, 99, right? He knew what he was not. I think Oracle is very good at that. They know what they're not. And that's why you see a company which transitioned from a $50 billion market cap in 2003 to $150 billion today. And I think VMware needs to pick and choose what they're really good at and what they're not. I mean, end user computing as an example is a massive chaos right now. They're trying to do everything under the sun, mobile and this and that. They just have to say, well, how do I take this revenue and make it into a billion-dollar business? Well, I don't see Simon Kreuzman here. So I have two more questions here. Is he here? Okay, he's talking to someone. We'll have one more question because it's good content. Open source, obviously, scale out open source has been really fantastic. No real commentary other than the fact that we love open source, third generation, whatever you want to call it. But we've seen Hadoop, we've seen Cassandra, these things are really rocking and rolling. So you got scale out open source, scale up proprietary. So you mentioned Oracle, right? So you've got different approaches, scale up proprietary and lock in this way or scale out here. So two questions. One is your take between the two different strategies. Can they coexist, yes or no? And then two, is the Hadoop era or the same as the Linux era? Can you go one-to-one matching? Is it different? I argued just so you know my position so you can either align with it or disagree with it. Linux had to fight Microsoft. You mentioned SQL Server. A billion dollars spent to kill Linux so they consolidated. I don't see anyone spending a billion dollars to kill Hadoop. So I don't think it's the same. That's my argument. Yeah, it's not the same. At the same time, it actually has more headwind simply because Linux was repeatable. You could do a million deployments of Linux and as long as the device driver guys were working on the drivers, Linux would become viral and especially the UNIX guys, the zealots who actually like the UNIX, they're like, well, I'm going to love this stuff, right? Hadoop is a lot more high-touch because it's transformational. It has to involve business logic and a lot of other stuff. PhDs, the code. PhDs, the comment code. And plus you can't have a million deployments of Hadoop here. In fact, it will be after, I mean, the market is global. A million same deployments, too. I mean, there's different flavors that you can connect here. Yeah, but at the same time, it's going to be global 2000 first and these are the guys who really want to analyze their own data. I mean, if you look at Fortune 5000 or Global 10000, they will actually use Salesforce's analytic reports and NetSuite and so on. So the SaaS guys will provide enough analytics and then you go to Omniture or Google Analytics for some of the other stuff. But the guys who really want to own their own data are the Global 2000 guys. Linux was ubiquitous. So can there be a red hat of Hadoop? It could. Although I think the way it's going to be different is if you look at the volume times deal size kind of product of these two, Linux was high in volume, was much smaller in terms of deal size than it came to Red Hat. I think here it could be the volumes alone, you know, the deal size itself could be because it's more complex stuff. And I think the cloud areas of the world have to really go and start to do that. I think provide the data model, the mapping the business use case to the data model, things like that. They have to start going and doing things. Yeah, and you get different platforms. You mentioned hardware is no longer a crutch and the ideal modern era. And I think the scale up and scale out, obviously mainframes are still a big business for IBM, right? I mean, and they continue to grow for a captive audience at 15, 20% rate for these guys. And IBM still knows how to continue to bless it and feed it and make. So I think the scale up hardware is here to stay. The question is what happens to the new applications and the next generation data centers and so on. And that's where distributed systems scale out. So you think coexistence is very viable? Absolutely. I mean, you know, I mean, we've been talking about paperless office since 1992. Paperless office still is elusive. Well, I agree. I would agree with you. And I think member of the old days client server interoperability was a big thing. Yeah, absolutely. Okay, we're getting the hook here. Okay, some great questions to some great answers to the point questions that we've been discussing at SiliconANGLE.com. So we're going to be right back with our next guest. Thanks for being on theCUBE. We'll be right back. Thank you.