 Here at EMC World, this is SiliconANGLE's exclusive coverage of EMC World, three days of wall-to-wall live interviews yesterday, we ended up day two, 45 interviews in the book, 19 more today. We're talking to everyone we can, executives, entrepreneurs, startups, ecosystem partners, customers, whoever has the action, we want to talk to them, we want to save them, we're going to extract that and share that with you. I'm John Furrier, founder of SiliconANGLE. I'm joining my co-host. Hi everybody, I'm Dave Vellante of Wikibon.org. Rich Napolitano is here, good friend of theCUBE. He's the president of the Unified Storage Division at EMC. Rich, welcome back to theCUBE for the, you know, millionth time. Yeah, good to see you again. There you go, John. All right, good. Yeah, so, you just came off your keynote, a lot of good feedback that we heard in the audience. You know, we were doing other stuff down here, but take us through what you told the customers. Yeah, so, feedback has been great, so it's hard to know when you're in the middle of it, but we want to take a different approach this time. We wanted to talk about kind of a vision for the future in a set of technologies. Not about product, but about a vision to enable your journey to the cloud. How do you transform from where you are today into the future? So, we talked about Viper and how that allows you to build out an agile infrastructure, heterogeneous agile infrastructure, both heterogeneous down into different storage systems and heterogeneous up into other management frameworks, OpenStack, Microsoft, et cetera, VMR, obviously. And then we talked about the implications of using fast software to meet the SLAs or your application by dynamically adapting the storage subsystem to fit the requirements of the storage of your applications. And then we talked about a new technology, which is in development and gave people a look at kind of where we're going. And how do you use multi-core and flash to produce millions of aisles a second at extremely low response time? And we kind of declared latency as the enemy. And then we continued and we said, you know, these things, Viper, software defined data center software, Viper software, fastest software, MCX, this technology around core scaling is software. And software has many properties. One of the properties software has is that it's malleable. You can add things to it. And so we demonstrated how malleable, where we took services, storage services, like virus scanning, and we ingested it into the platform so you don't have a separate appliance. You can ingest that service and extend the value for opposition of the platform. We did the same with RecoverPoint. We took RecoverPoint that generally runs on a separate discrete appliance and ingested it into a VNX and demonstrated RecoverPoint inside the platform. And then we said, you know, these things are all software, right? So, you know, Viper, fast, MCX, this flexibility, but software also can be deployed in many ways. You know, I happen to use a droid and droid runs on phones and runs on tablets, same with iOS. And so you can deploy your software in many ways. You can deploy your software in purpose-built platforms like a VNX or VMAX, but you could deploy your storage software in ESX. So what we demonstrated on stage was a virtual VNX. We took the VNX stack and put it on top of VNX and delivered SIFS and NFS services on stage. And then the last piece was, we said, you know, you can even take that same software and push it up to the cloud. So we pushed it, we worked with one of our partners, Verizon, in this case, we could do with AT&T or others, Rackspace, and we took a virtual VNX and instantiated it in the cloud. So that was a long answer. That was exactly what we covered. You just gave us your whole keynote in three minutes. It was perfect. So help us understand this whole software-defined thing and sort of where you guys fit. We had John Rose on the other day, I think it was yesterday, and he talked about how, you know, all this function lives in middleware, it's very tightly tied to the controller and sort of how Viper begins to change all that. Help us understand where you fit in that hole. So Viper is this horizontal layer. You guys have this, we talked to Brian Gallagher about this. You guys have this hardened stack, you know, that's been built up over, you know, a decade plus. So how does that hardened stack fit in to this horizontal layer? Sure, that's a great question. We cover some of that in the keynote, but I think it's important to refine that message and continue to repeat it because it's a little confusing. It's complicated for people, yeah. So, you know, when we think about Viper, you need to think about Viper in two ways. First, it's control path functions, which is federation of management. In other words, one of the challenges, you know, one of our core beliefs in EMC is you have the right tool for the right job, the right technology for the right use case. And so we have a portfolio of products, and that's always been in our DNA. But the challenge with that is each one of these is managed discreetly and separately. So Viper represents an aggregation of the APIs across these platforms. That's very, very important for our customers to make it easy to deploy all of our technology. So that's a control path function. Now, on top of that, you can build other services like billing and metering and provisioning and simple provisioning, which can drive consistency of that experience down for our users across our platforms. That has huge value for our customers. Now, since what Viper is doing is exposing control interfaces up based on RESTful interfaces, you can tie in our Viper management now into other frameworks like OpenStack and VMware as well as Hyper-V, Microsoft and Environments. So that's powerful in terms of the control path side. Now, when you cross, so you can build services on top of that, billing, metering, provisioning. And then on top of that, you can have some things in the data path. Now, the things that make sense. The things that are, in fact, latency insensitive, like an object store. So build an object key value pair architecture, implementation that has its roots in Atmos and in Centera, and allow that object store to be built upon physical storage systems because Viper has no physical assets. It's an abstraction on the control path and then in the object side. So the data's got to land on some disk somewhere. Now we happen to have a lot of great, amazing disk subsystems that have different properties, price, performance, availability characteristics, scalability, et cetera. So Viper overlays that value proposition. So sometimes Viper will be in the data path. Most of the time it won't be for more traditional applications like file-oriented applications and block. And so it allows you to aggregate these systems and build on top of the fabric of what's underneath here today. So they really are very different abstractions. So you'll access those lightweight services that you mentioned before. I mean, not to integrate billing and provisioning, things like that. But then you'll access the southbound services directly through your protocols. I mean, and we'll expose those interfaces so other people can plug into that framework, the Viper framework. And it allows you to solve our original, sometimes I call it our original SIN, which is that we have many of these and we need to basically aggregate them in a way that makes the experience across our portfolio easier to consume. So what's your penance? Is obviously Viper's coding is part of that penance. What do you have to do to play in that more? We need to just invite more consistency and make it easier for Viper to move forward. So Gallagher and I in particular, when you think about, as we talk about the technologies, fast de-duplication and compression, we're architecting them more and more similarly. So the experience between a product to be more similar. Isilon is thinking about building out notions like fast cash. And so as they articulate those things, their interfaces need to be more and more consistent. It just allows us to innovate more rapidly and bring that value all the way up to step. So the criticism you're going to get, Joe can talk about open, he can talk about choice, et cetera, et cetera, but the criticism you're going to get is, oh well, Rich, this is proprietary. So the answer to that obviously is going to be, okay, well, if you can deliver a function that's greater than when I can get an open source, then that's okay, our clients will pay for that. So can you do that? How long of a leave do you have in doing that and is that sustainable? So it's important for us to have these frameworks that are open because customers have mixed environments. So you need to embrace that. And at the end of the day, you look at the rate of innovation in our subsystems, I mean, it's relentless, it's just relentless. And that's the game, right? I mean, at the end of the day, I look at my roadmaps and I have things, five years out that we're working on that are going to take that long to mature. And so, and you need to make those investments now. I mean, some of the stuff that we're talking about now, we started five years ago. And some of this is very, very complex system stuff. And that's the game, right? The game is technology is our friend. Technology allows us to attack cost and complexity in the infrastructure if you're smart about it. And so we got to keep going. We can't rest on all walls. Rich, I got to ask you on your keynote, we'll look at some nice little graphics from the blueprints that they do, the drawings. But you mentioned a software defined data center in your keynote. One of the things that's been mentioned is OpenStack. You mentioned that. So, David and I have been talking about this open choice message that you guys have. And it's just too fuzzy, right? Back in the 80s, 90s, it was multi-vendor. It was a nice clean message. Multi-vendor meant if you have a vendor, you can go with other vendors and things will work through some standards. It's evolving right now in this market. So what is the new multi-vendor? If multi-vendor was boxes, this works with that box, that port connects, this network works with that. The lower end of the stack. Now you got software defined data center, you have open source. And we were talking yesterday that open source is kind of like the new multi-vendor. So you got to play with the open source communities. You got OpenStack. Is that the new open? Is that the new choice? Is it like, hey, it's not about open source per se. It's just about working with open source. How do you view that as you lead your engineers and developers? So it's all opportunity right at the end of the day. So if you think about a storage subsystem, we're agnostic to the host. So in the old days, it would be VMS and Solaris and AIX. And so now those words have changed. It's OpenStack, it's Hyper-V, it's VMware. What's any different, really? And so we need to continue to plug into the physical infrastructures, traditional physical operating systems and those virtual infrastructures. What's really changed? I'm agnostic to that. My business, more than 50% of my business is connected to Hyper-V. That's a fact. We're driving more and more into the service providers for all of our products. We announced VMAX Service Provider Edition. There'll be a VNX Service Provider Edition, which means we need to be in OpenStack because a lot of those guys are going to go that way. So at the end of the day, that's what the new multi-vendor is. Is what, the environment? So is it a software? It's the software that defines those environments. In the old days, it was the server vendor architecture and there were more choices in the old days. You had power, you had alpha, you had 20 different Spark, right? 20 different architectures and then they ran different operating systems. Now it's basically Intel and you have physical operating systems and virtual operating systems and then you have a set of control path functions. We all hope that SMIS was going to solve these problems. It didn't. Amitabh was on yesterday talking about this and his comment was, you don't have to be open source to be open. That's true. And so, what you're referring to is, yeah, you can play in open source as a contribution so the communities of open source are a good benchmark to balance. I know I like that Allen Top guy. I know I like him. He's solid, yeah. He's good guy. He's one of the family. Yeah, John Rose is impressive too. He was pretty impressive. And he talked about the service provider, Mark. He's saying, hey, that stack, you can't just move that across the enterprise. That's right. And he was talking more about those components. Does that affect your business or are you still plugging into the normal? So we'll plug in to all these management frameworks. And we have seen forever and for all time the notion that, quote, storage is a commodity. And what people mean by that when you really unravel it is the disk is a commodity. What we do is hard. It's just hard. And the beautiful, and when I hire somebody in my office I have this disk drive, because I'm an old guy. And I hold this disk drive up and I say. 10 meg. And I hold this disk drive up and I go, you know why we have a business? And I'm like, why? Inquisitive luck. I go, because this thing doesn't work. It's disk drive. And you know what, it never worked. And the day it's born, it slowly decays. And our whole existence as a business has been, how do we ensure that your data is protected from this underlying substrate that doesn't work? And I say to you, my whole career has been defined by this disk drive. But your career is defined by this flash drive. And the good news for you is it works worse. So. It decays from the tradition. It's like, and when it does decay, it goes, so there you have a long career ensuring people's data is protected. On the career side, let me ask you that, because obviously flash has changed in the game. It's from a software standpoint. So I mean, you've been directing engineers. You've been doing a lot of organic R&D and innovation on your side as well as doing some good M&A stuff. But so as you hire new software engineers, Meolden was a pretty classic software development. Now, and disk drives were working for the latency issues of memory. Now it's flipped around. Discs and memory have this symbiotic relationship with flash. So does that change the hiring? It's system design. System design, right? I mean, in fact, our internal word for it, our internal group, it's probably not outside my group. We call them stone mason, right? And why is that? So if you go to Europe or, you know, any place in the East Coast or whatever, you have these old churches. If you tried to build a church today, you couldn't do it. And they worked on these projects for hundreds of years and they were works of art. And so we're looking for stone mason. So if you're a stone mason, we're looking for you. And what we mean by that is great system architects that see the entire system in their head and understand, you know, principles of latency and architecture and NUMA and memory management. And, you know, it's basically operating system internals. When you look at my team, they're what I was 30 years ago, operating system internal developer. Yeah, systems management's coming back. And I mean, basically it's an operating system. It's an operating system. You got it. It's the same skills. Okay, we're here inside the Cube, Richie and Paletano. President and friend of the Cube. Always great insight. Obviously had keynotes online. You can always get the replay on EMC TV. Here in the Cube, we just go a little bit deeper, a little bit more casual, nearly kind of distracting out the signal from the noise. I'm here with Dave Vellante. We'll be right back with our next guest after this short break, Silicon Angles. Exclusive coverage of EMC World Day Three. We'll be right back. If you're a stone mason, we want you. That was great. Good job.