 From Las Vegas, it's theCUBE. Covering EMC World 2016. Brought to you by EMC. Now, here are your hosts, Stu Miniman and Brian Graceley. Welcome back to Las Vegas, here for EMC World 2016. Getting towards the end of day two of three of our coverage, I'm Stu Miniman joined by Brian Graceley. Happy to welcome them back to the program. Manavir Das is the Senior Vice President of the EMC Advanced Software Division. Welcome back to the program. And, you know, gotta give us the update. You know, of course, ECS, everything going on. Mentioning the keynotes. What's new for you in the last year? Yeah, it's been a very exciting year. New releases of ECS, a lot of momentum. Also, we've released a new product. We extract Neutrino that I'm very excited about. It's very much based in open source. And then new updates to Viper as well. So a lot going on. Yeah, so we'll definitely want to unpack those. Maybe, let's start with Viper. I know David Golden talked a little bit about the keynote on Monday. But, you know, where are we? The whole software-defined storage space has been evolving, you know. So I think a year ago, we made a decision to open source Viper controller as a new project called Copperhead. And over the past year, we've built the community up. We've had folks like Intel, Oregon State University, et cetera, really jump on. And what we've done is we've literally done it in a way where my engineering team doesn't even have a code repository of their own at EMC anymore. It's all on GitHub. We developed there. The community contributes there. And perhaps the most exciting thing is we just announced yesterday Viper 3.0, a new release of Viper. And literally, it has come from that GitHub repository and the main feature we've done, which is a new Southbound SDK, by which any vendor can add support for a new storage array. That was actually conceived and done by the community and contributed into Copperhead. And we've picked that up and have provided that as part of Viper now. So exciting times for Viper. Yeah, totally different engineering model than EMC's used to. Totally different, yeah, yeah. So we take snaps every so often. We do a lot of QE to get it ready for enterprise customers and then we make it available as the next release of Viper. But there's really no difference between Copperhead and Viper in terms of the code base, yeah. So can you talk a little bit very different, as Brian said, there's the, I put it all out there and gosh, somebody else might monetize it and run away with it versus kind of bringing it into house. How do you guys balance the development? Yeah, I think the simple principle we've always followed that's worked well for us at EMC is if you take the customers where they want to go, it'll work out in the long term, right? And this was a clear ask to us from the customers to have this open source to really be an asset of the community. And so if you look at the decisions we actually made, too, but the way we did the licensing, all of that, we've done it as open as could possibly be, right? So yes, somebody could take this and monetize it tomorrow and good for them, you know? That's the way we look at it. So, you know, with the open source solution, you can usually see how many contributors there are, how many downloads there are. Yes. How's the progress, you know? How much revenue is there? You know, customer deployments, you know, production, things like that. Yeah, I think, and really the customer deployments is the interesting one. We've seen a lot of traction now with enterprise customers where they have started standardizing on Viper as the software that they use to orchestrate all their storage. And in fact, what really drove the Copperhead project was that we had some customers who got to the point where when they were doing an RFP for new storage, they had a requirement for the vendor that you need to be qualified and certified with Viper. And that's why we did this whole effort to open it up and do the Southburn SDK. So I think we've seen a lot of traction. We're very excited about it. And we've got it now. Every array you get from EMC, certainly, you get Viper with that and you can use it to manage your storage. Yeah, so taking open source in a little bit different direction, we were both at OpenStack Summit last week at Austin. You announced the Neutrino notes this week. Tell us a little about that and what's going on in the OpenStack world. So I think one of the most exciting things that's happening in IT really is that in the community have sprung up these new infrastructures and frameworks, whether it's OpenStack was really the bleeding edge, Hadoop, now we have Mezos, we have Kubernetes, all of these efforts, right? And you see customers beginning to use these things and play with them, but they're not quite ready for primetime. A big reason is because the storage that's linked with these is not quite ready for primetime. And then just in the whole operational and consumption experience, it's not turnkey. So what we decided to do at EMC was we said, let's really take the customer where the customer wants to go and let's give them a turnkey experience for open source infrastructure, right? And that's what Viper Neutrino is really about. It's like to use the term very loosely, could we do the iPhone for platform three, right? That make it really easy to use. You plug it in, turn it on, you've got OpenStack deployed, there's a new version, you want to do an upgrade, it's all automated. So that's Viper Neutrino. The first, the start of it is OpenStack. So what we have now in the box is turnkey OpenStack, you plug it in, you've got a whole OpenStack environment. It looks and feels like normal OpenStack, all the interfaces, everything. But on the covers, we've made it robust. And then from here, we go forward, we do new services within that for big data, for other kinds of compute platforms as well. Yeah, I know when we were, like I said, we were at OpenStack some, they're making progress, but people are still saying it's difficult to operate, it's difficult to upgrade. You guys addressed, you tried to really tackle that challenge. We didn't address that, we're saying, okay, let's reduce the barrier for customers so they can really discover whether this is for them or not. Now, of course, I am from EMC, I'm a big fan of what we build with VMware, and as you know, VMware itself is working on new products that also take the same approach, and we will support them very heavily in Neutrino, but we're really giving the customer choice, right? That there's a variety of models out there, regardless of which way the customer wants to go, let's make it a very robust turnkey experience for them. Excellent, excellent. One of the big announcements this week was around the virtual stream. And on the back end, or sort of behind the scenes, is ECS. Talk about not only what virtual stream is doing, but service providers trying to offer Amazon-like services or S3-like services. Right, and that's a great use case for ECS. We have Atos in Europe that is a service provider that has built a large ECS environment that they're sharing with other customers, including Siemens, that'll be on stage tomorrow with Jeremy Burton talking about that. I think what they really value in ECS is that with the product in hand, service providers can offer the customer not just an S3-like model, but also a more dedicated, a hybrid model where the customers have some equipment on-premise, some equipment in the service provider's data center, right, and it opens up some opportunities that you cannot get from a pure public cloud. For instance, if the customer has particular equipment that they want to place next to their cloud storage in the same data center, how do you go to an Amazon or an Azure and say, I want to roll in my equipment into your data center, it's not that easy, but with a service provider, like an Atos or a dimension data, you can do that kind of thing, right? With VirtualStream specifically, we've worked with them for quite some time. We've, what people don't know is that we've had an exabyte scale storage footprint with VirtualStream together for quite some time, and what we really announced yesterday is now making that more generally available, right? So we're very excited about that, and I think my message to the customer is, you know, ECS is great technology, and now you can consume it in every way you want. It's a software product, you can run it on your own servers, you can take an appliance and rack and stack it, you can go to VirtualStream and consume it in an off-prem way, where you don't have to manage it, right, and it's your choice. Yeah, so many of you, you know, we find customers that we talk to, it's the multi-cloud world they're living in, so you talk about the on-prem, the off-prem, you know, service providers, public cloud, we tend to find it's, you know, yes, all of the above, talk about how you kind of plug into frameworks and management tools that are out there. Yeah, I think generally as a company, we strongly believe in that, in the customer's choice for cloud. So if you look at the work we've done in the last year on our arrays, right, one of the things we've done is, we've built tiering models from a number of our arrays, where they can automatically push the cold data into the cloud, and of course, we've built support for ECS and for VirtualStream, but we've alongside built first-class support for other cloud vendors as well. So for instance, you can take an Icelon and do cloud pool steering and push the data to Amazon or Azure, right? Just as easily as you could to an ECS that you have or to VirtualStream, because I think that is important for the customer, right? To have that choice, and I think the other part of it is with respect to frameworks, all the APIs we're doing are completely standard, right? So we adhere to the S3 protocols, the Swift protocols and all of that, because that allows this ecosystem of whether it is gateways on the top or data movers and the like, to work 100% with our product so that the customer doesn't get locked in in any way with our technology, right? Well, what do you say, the software-defined storage has been around for a little while, and you're obviously right in the heart of it. What are you seeing in terms of uptick, people understanding how to operate these sort of software-defined environments? I think that's a great question, because these two words you said, uptick and operate are the two keywords. I see a lot of uptick. Most of the big customers I talked to today, the first thing they want to know is, can I consume this as software, because they're all picking hardware vendors and standardizing their hardware across the data center, right? And the prime advantage of the software model is I get elasticity in my own data center, because I can deploy the software in 10 servers, and if I find the need for more capacity, I deploy the software in 10 more servers and I double my footprint, which is very hard to do with an appliance model, right? So that's what's attractive, so we do see uptick, but what we're finding at the same time, Brian, is that these same customers are used to a certain turnkey experience with the appliance, right? There's a single number to call, they take care of everything from the hardware to the software, fix the problem, and in our deployments, like we have customers like Verizon, has a very large deployment of ECS in a pure software-defined model and ScaleIO, and when we get on the phone with them, it's typically about that issue, that there's something going on between the fact that they've got their own hardware, and they've got CoreOS, and they've got Mezos orchestrating things, and then there's our software, and it's a little less turnkey because the integration is not built into an appliance. So I think that's the journey for the customers, and I think it's not ready for prime time in the sense that you have to have an engineering team as a customer to really go that way, but I think the payoff is certainly there, and we're seeing that. Can you give us the update? Where are we with kind of the financial models of consuming this between kind of the CapEx or like Public Cloud is more of an operational model? That's right, and I think certainly our traditional model with equipment has been that customers purchase that upfront and in a certain time period. I think EMC, even with our managed services model, has been quite open to a utility model, and you see that shift. So for instance, we talked about Neutrino, which we just launched, right? For Neutrino, we have put in a new model for this product where we sell the hardware and the software separately, and we sell the hardware on a cost plus model. So for the customer, we say here's the actual sheet of the cost of the hardware. You can compare it with going yourself anywhere and getting the hardware, and we have a small percentage premium on that because we're doing some work here, and then the software is on a utility model, right? So that's sort of how we are, and again, it's a matter of, the customer's always right, right? And we always want to take the customer where they want to go, so I think, yeah. Okay, so, one of your, really appreciate coming back, giving us the update, lots of changes happening in the software-defined storage world. Wrapping up, day two of three of theCUBE covered. Check out siliconangle.tv for all of the videos from EMC world as well as upcoming events. Thanks for watching theCUBE. Looking back at the...