 The Cube at EMC World 2014 is brought to you by EMC. Redefine VCE, innovating the world's first converged infrastructure solution for private cloud computing. Brocade, say goodbye to the status quo and hello to Brocade. Okay, welcome back everyone here live at EMC World in Las Vegas. This is the Cube, our flagship program where we go out to the events and extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE. I'm Joe Mike, co-host, Dave Vellante, co-founder of wikibon.org. Our next guest is Sal DiSimone, VP and CTO of EMC's Advanced Software Division. Welcome to the Cube. Thank you, happy to be here. Glad to be back. Great to see you, welcome back. And obviously Joe Chuchy put a lot of pressure on you. He said the software driven enterprise, software led enterprise is all about the future. So what do you mean by that? Just want you to give us your definition of software led, software defined. I mean, part of what we're recognizing is the fundamental changes and shifts to the way people want to consume, deploy and use IT. And so a big key to that is to be able to bring the convenience and the simplicity and the economics that people like with public cloud but be able to deliver it in such a way that people can have it on their own premise in full control under their security. So kind of bring the best of both worlds. And a key enabler to all of that is to be able to deliver this infrastructure in a much more efficient than agile manner. And that leads towards software based solutions where more of the configuration, more of the function is in software that's easy to configure and easy to tune and less of it is in super, super specialized. Explain to the folks, what does the Advanced Software Division do at EMC? Because Joe Chuchy also explained the innovation strategy was M&A and organic R&D. Are you in that R&D piece? Are you by the product portfolio? What, just to quickly explain the Advanced Software Division. So it's actually a combination. So the Advanced Software Divisions Charter within EMC is to define the software defined strategy for EMC. We have several products that fall under our division. We have obviously Viper, the whole software defined storage portfolio. We have the new elastic cloud appliance, storage appliance. We also have EMC's existing object storage systems Atmos and Centera. And all of our storage resource management and network and data center management products all brought together to kind of help solve the problem. So I got to ask kind of a, I guess a generic open but I guess complicated question is EMC talks about best of breed, nice product portfolio low end, mid range, flagship, the normal stuff, but best of breed is what Goulden talks about. But yet scale out open source is disrupting horizontally those known silos. How do you look at that disruptive force? Obviously commodity hardware is changing the game. We heard Jeremy Burton saying, hey, we think software is going to be all the action. So you're in the software advanced group. You look at that trend coming down main street tech land. What do you, how do you guys prepare? What are you guys doing differently? So first of all, we think we're really well positioned to be successful in this software defined world. We've actually started a lot of these efforts several years ago. So it's not like we just woke up yesterday and are trying to kind of figure out what to do. So I don't think the open source, the software on commodity necessarily changes any of the best of breed because at the end of the day, you still need- So you're saying best of breed can exist, co-exist with the disruption? Even within the open source software on commodity marketplace, there's going to be better implementations than others with different characteristics, different price points, simpler to manage, harder to manage, higher performance, lower performance. When you're going to run different workloads, you need different characteristics in your storage system. So the fundamental physics aren't changing with software and the need to build systems that are engineered well is not changing. So even within the software world, there's going to be a large variety of software systems optimized for different price points for different workloads. So in that regard, it's not really fundamentally changing. So I was talking to a big bank the other day with some practitioners I was speaking to, gave my speech, a guy comes up to me afterwards, huge bank, one of the top banks in the world, and said, I want to talk to you about Viper. He said, oh, cool, great. I always want to talk to a customer about some new stuff. And he said, what I really want is an open source object store that really goes fast. He goes, but I can't get one. So really looking at, we started looking at Viper and it looks good. Some things, we're not quite there, but now 2.0 has been announced. And then I said, okay, what's driving it? And he said two things. One is simplicity and the other is basically simplicity. So we think this is the future is what he said. But at the same time, we want a way to consolidate all the different platforms that we have in the organization. Are those the two big drivers that you're seeing? Is there one that outweighs the other? Is this the future of the storage business? Is this a way to consolidate the stovepipes or is it both? Both. And if we go back to last year when we announced the Viper strategy, there were really two different objectives that we had. One was for people within existing storage estate, let's provide mechanisms that help them operate that in a much more efficient way. Essentially allow them to apply a cloud operating model to the gear they have. It's not about buying a new storage system, not about fork lifting, it's the stuff you have. Let's give you ways to operate that more efficiently from a capacity, from a migration, from a provisioning. So that's one aspect which touches. The second thing that we're trying to solve with Viper though also is to build the next generation storage stack. Designed from the ground up with these cloud and third platform properties, built in software, designed to be elastic, designed to be scalable, designed to do this and not sacrifice performance, not sacrifice geo protection and these other capabilities. And actually do those in a way that we provide our customers a bridge from where they are today to where they want to get to. And I think EMC as a company is uniquely positioned to actually bring customers along. So Sal, so help me understand this. So you've got some, you're building a stack, right? And it takes a long time to build a stack and get it functional and robust. But you've got some native, HDFS for example, you know, object store. And then as well you accommodate existing EMC and other platforms. So we're going to have Brian Gallagher on tomorrow but I remember he answered this question at one point and he's very open. You know, he's like embraces Viper. I'm like, aren't you worried that, you know, this thing, he said, his answer was my job is to make sure that my stack keeps, you know, outpacing the rest of the industry, be more robust. And your job is to catch up to him in a way. Isn't it? It's on the software side. It's to accommodate him, but at the same time, yeah, but that's my question is, are you building out a stack of comparable functionality that might take a decade, that ultimately will be pure software? Or should we think about this differently, that no, you're going to always have to have a box to get that function, that latest, greatest function? So I think, let me answer it in two different ways. So again, going back to the kind of fundamental architectural separation of the control plane from the data plane, which is one of these fundamental principles that we came with Viper. There's a lot of storage systems that have capabilities that are unique to those that serve a very specific purpose and are tailor made. If you have enterprise apps with relational databases and you want massive throughput and you want a tank that will never get, Vmax is what you should do. And it'll be a long time before any kind of software based system running on random commodity can compete to that with those characteristics. Now there's other workloads that you don't need the tank and you don't need that you could do with something. So there'll always be this mixture of platforms with important characteristics. So for customers with a corpus of existing storage on existing arrays, and as those arrays evolve in the future, we have Viper as a control plane that can bring automation. I mean Vmax is not the simplest device to manage. It's got a lot of sophisticated features, but it's fairly complex. So in a Vmax world, Viper brings a significant amount of automation and plugging into the various cloud stacks that makes Vmax sort of simpler to operate. So there's a lot of value Viper brings into Vmax. On the software side, we are trying to build out the next generation. There are workloads where those stacks are useful today. These stacks will evolve. The data services in those stacks will evolve over time and we'll see how it plays out. But in the Goulden workload slide, what you just described covers a lot of that space. Your TAM is pretty damn big. Isn't it? Yeah, yeah. Okay, so you mentioned Vmax as an example. That's kind of an edge application for you guys. What else don't you touch? What else won't you drive a truck through in the next 10 years? It seems to me like you're very well positioned to knock down a lot of those workloads. Yeah, I mean, we want to be the state of the art best software based stack that drives EMC into the next generation. So I want to make sure I'm thinking about this right. You separate the control plane, the data plane, which is the fundamental concept that you put forth last year at EMC world. But then within the data plane, there's still a strong connection today between the hardware and the stack. Right, but now as we grow and mature our software stack, and that the fact that we now support running the software stack on commodity hardware, we're starting to expand the choice that customers have. And that is the future of storage as you see it. Is that fair? Yeah. I don't want to put words in your mouth, but essentially that stack running on commodity hardware. Awesome. Okay, so playing through me to just between scale IO, vSAN, and Viper real quick. Because Viper's more of a broker across the platform, right? I'm just trying to, you brought up that earlier. So like, how do I, the customer wants to look at scale IO, vSAN, and Viper? Let's take scale IO and Viper. So again, Viper is two different pieces. There's a Viper control plane, which is all the automation orchestration. There's the Viper data plane, which we had originally object in HDFS data services. There's now Viper block software based data service, which is scale IO. Scale IO is the implementation of Viper block data service. Scale IO is a block implementation in Viper. In Viper, for software data plane. Now scale IO and vSAN are really fundamentally trying to solve the same problem, which is how do you build a server-based SAN? Take a bunch of servers with disks, ethernet connections, and basically pool and aggregate these, and turn these Linux nodes with disks into, or not Linux in the case of vSAN, if the effects, but turn commodity nodes into a shared pool. So at some fundamental conceptual level, they're the same. And from an implementation, they're very different because the starting points were different. With vSAN, it's very ESX, vCenter focused, very tight coupling and integration into the VMware stack, which is very nice. The scale IO Viper block point, we're looking at storage as a horizontal. So this kind of server-SAN, not just for VMware, but for VMware. So can you geo-distribute HDFS and block and protect it? Yes. And object. And object. And object, all right. That was a good, yes, and I think. So also object, HDFS, geo-block. And block, all geo-distributed and protected. Yes. On Viper. Yes. Now, the interesting thing about geo is a whole topic. We can talk about it. But when we look at the innovations and the truckload of patents that we're filing around Viper, our geo-implementation, we feel, is state-of-the-art and kind of leapfrogs, where any other comparable solution is. Because with geo, there's always two competing forces. Big trade-offs. Big trade-offs, always, right? It's like, do I have local access to data for fast access? Yeah. But then how much storage capacity am I using globally to achieve that, right? So you have schemes on one end, I have five sites, I have five copies of the data. In each site, I get very fast local access, but it's 5x overhead. There's the other extreme where you have geo-erasure-coding-type schemes where- Very efficient. Very efficient, but I got to stitch together these objects in slow. And so some of the breakthroughs we've made in Viper is to kind of bring the best to both worlds. And we have some unique intellectual property that we've developed, which uses a combination of local erasure-coding and geo-compaction algorithms, which are brand new, which allow us to do, for say, five sites, the overall storage overhead, 1.67, so close to erasure-coding, yet every site has a full local copy of its data, no WAN traffic. So that's really amazing that we can do those two things. And this is one of the reasons when you asked about, well, you know, storage software on commodity, everything's equal, everything's not equal, right? Because we have some of these capabilities that are unique intellectual property that we bring that other software-based things running on commodity stuff. So you can still differentiate in this open-scale-out commodity gear. Absolutely. So clients want to play with some scale-out low-cost gear you guys can drop into, like, building block. We can do that, and we can provide them with geo-distributed, active-active, object, HDFS access to the same, I mean, there's a whole bunch of capabilities that we bring to the table that are unique. The object in HDFS to geo is pretty cool. To geo-distributed. It's very cool. We got to talk about the cloud in the box, the elastic cloud storage appliance. Right. You got software division. You sell them boxes? So we can't help ourselves, right? No, but actually, if you- Is it a box as a service? Is it in the cloud? No, no, it's a box. Okay. You can install it. So you can duplicate the cloud. So we have a whole spectrum of customers who want to run storage software on commodity. Some want to buy software, buy their own hardware, and build things. Others are not interested in building science projects. They want the convenience and the TCO of software on commodity, but they would prefer to buy a system. So what we've done is give people choice. It's the same Viper software running on commodity hardware. Your choice, whether it's on ECS appliances or your own, you can mix and match in a given deployment. So we're just giving customers more and more choice. And is it really, what is it, 28 to 30% cheaper than public cloud? That's a point in time. Is that 2G math? No, no, no, this is actually NG math. And when we do those computations, it isn't just the capital cost, but it also brings into what's your power cost, what's your admin cost, what's your cooling cost. So it's a really a TCO measurement because when people go to public cloud, they don't need data centers, they don't need- It's a service, it's a line item. So we're comparing apples to apples in that regard and we are highly competitive. And from a feature set, we have a differentiated set of data servers, active, active, GEO, all these functions, which really have our stack leapfrog what these public cloud- And the demand, am I right, is coming from organizations that essentially want to duplicate public cloud capabilities, flexibility, economics- Public cloud economics and convenience. But they don't want to go off-site. Yeah, they want choice. And I think we're beginning to offer some solutions that give people those choices. So we're very excited. Thanks for coming on theCUBE. We really appreciate it. I want to give you the final word. Share with the folks out there why this year is so important for you guys. What's going on right now that is such a game changer and why is it so important? I mean, so from an overall EMC perspective, we came out last year and announced Viper as our strategy. We hadn't announced any product yet. So there's a lot of skepticism. Well, as EMC is a hardware guys. So we went through last year, we released Viper when we said the first version. Now we're coming out with the second version. And I think this is really important because as we hit each of these releases, customers begin to triangulate. Not just what we say, but what we're actually delivering. And as we get multiple releases with greater and greater expanded functionality, hey, people can start to draw this curve and triangulate and say, hey, these guys might actually pull it off. So I think this is a very important year for us to bring these new products that are highly differentiated, very strong TCO, and basically prove to the marketplace that the EMC hardware guys really know what the heck they're doing when it comes to storage software. Well, Sal, CTO, EMC Advanced, so thank you for coming to us. This is theCUBE. We'll be right back with our next guest live in Las Vegas for EMC World 2014. We'll be right back.