 Live from Las Vegas, Nevada. It's theCUBE, covering EMC World 2015. Brought to you by EMC, Brocade, and VCE. Hey, welcome back, everyone. We are live at EMC World 2015. This is theCUBE. I'm John Flurry. My co is Dave Vellante. Our next guest is Josh Goldstein, Marketing and VP of Product Management at EMC. Extreme, I welcome back to theCUBE. Yeah, thank you very much, John. Great to be back here again. A lot's changed. So you're doing what? 50% of overall revenue now for EMC? Oh, man, I don't know how much overall it is, but yeah, it's growing every day. Really great numbers. We saw the guy's projection really kicking up. What's been the big factor? I mean, obviously, and what's changed since our last year's interview in terms of product, customers, features, road map, share some data? Yeah, so last year was a tremendous year for us. The product came into the market late in 2013. So 2014 was our first full year. We executed, unbelievably, became the number one all flash array out there by quite a wide margin. We put out a bunch of new releases, put in some incredible capabilities in the array, so we enhanced our data reduction. We brought out some really sophisticated features around snapshots and copy data management. We're on the cusp of our 4.0 release that was announced here at EMC World that's going to up the ante all over again. So a tremendous amount of work going into the platform. What was the biggest thing you learned from last year? What magnified learning could you share? I think one of the things we learned is that this is kind of something that customers have found insatiable. You would think that people, I guess the common prevailing wisdom on this is that you're going to apply flash for a hot workload. That's what everybody thinks. I got a database that doesn't work well. Let me put a flash array on it and then it'll run nice. That's all true, but what's really been amazing is how fast customers are landing and expanding. They're seeing that work and then the flash array tends to just, cover up workloads from all over the data center and it changes the way operations work for companies in terms of this. What's been the impact of the portfolio? Because what that means is you got more sales than you thought. Yeah, Extreme IO grew very quickly. Fastest growing product we've ever had at EMC. And I'd say for the portfolio, it's obviously great to have a real winner like this knocking it out of the park. But it's also great because we don't talk about it a lot but there's a lot of pull when we are doing Extreme IO deals or other things that EMC has in the portfolio. So it's been tremendously successful in V blocks as an example, pulling along the whole converged infrastructure with Extreme IO. We've done a lot of work with V specs. We've got Extreme IO integrated to practically everything else in the portfolio. It's pulling along Viper. It's pulling along Vplex, RecoverPoint. So there's a lot of things that have happened in the ecosystem beyond just the core Extreme IO array. Do you feel you're energizing other parts of EMC because when you have that kind of locomotive power with Extreme IO, it tips the scales a little bit on balancing architectures. What's your product view there? I mean, has it been an enabling, energizing or frightening? Well, no, I think maybe a little bit of everything. I don't know. But I think- I asked kind of the same question, by the way. We'll see if you're- I think we've all found it pretty energizing. Certainly for the Extreme IO team, this has been an amazing experience. And for EMC, EMC is on the forefront of this amazing technology transition that's happening in the marketplace. It's pretty rare that you see the incumbent vendor, the one that's already leading in an existing technology, also become the leader in new technology. And EMC is better at that than anybody. So I think that's what's energizing for people. So tell us about 4.0. What's that all about? We always talk about the stack, how most all flash arrays have a less mature stack than a Vmax that's been around forever. What's 4.0 about? So 4.0, I call it the no excuses release because if a customer's been sitting on the sidelines wondering if an all flash array is mature enough for them, this definitively says, everything has been addressed, the scale is there, the data protection is there, the ability to do replication is there, everything that you could possibly want is in the platform now. And so it is really the time for this to go mainstream. And as big as we are now and as much as we grew last year, I think 2015 is going to really be even more amazing because of the things that are coming in 4.0. You know, Josh, the interesting thing about flash, very lately, and you guys were kind of, I think the first to push on this is everybody obviously to say flash performance, great, get it, put my database on there. We were just talking to Callaway and they were talking about using snapshots, how it's going to affect data sharing and leveraging copy services to improve developer productivity. That's a massive trend that you've seen. Can you talk about that a little bit? Yeah, yeah, so what are customers are really able to do with the Extreme IOS platform? Everybody knows that a flash array is going to go fast, it's going to have low latency. And then the next step was, we need to put in some data reduction features so that we can fix the economics of flash. We can take a more expensive media and effectively make it competitive with disk architectures. But the next frontier that we're really pushing in Extreme IOS technology is copy data management. It's this idea that when you look across an enterprise data center, any big application stack, Oracle, SQL server, Epic databases if you're a healthcare provider, SAP, SAS, all of these things, you have your production instance that you run that's where you normally would think about using a flash array, but you also have a lot of copies of that production data. And creating those copies, managing those copies, refreshing them, having those copies be high performance, that is a big challenge in the operations of how data centers work. We saw huge opportunities to improve that whole workflow with Extreme IOS. So for us, it's not just about using flash to fix a production performance problem. It's about how do we enable a workflow end to end over the whole life cycle of that application. That's what's been going into the product. There's a lot of new enhancements in 4.0 around this as well. So you get a lot of tailwind from the economic factors where Djupe compression and things like that, as you said, address the higher cost of flash. But as well, if you can start combining that with reducing the number of copies, forgetting about the labor content for a minute, the CAPEX to support an application is going to be lower. Now we know that flash is essentially lower cost now than high performance spinning disk, if that's even a, that's an oxymoron, I guess. But it appears that flash is potentially positioning to be less expensive than any spinning disk. Do you buy that theory? I actually buy that, and there's been research. You guys did some phenomenal research in this area where you looked at, in a typical enterprise application stack, how many copies of the data exist? And what's the potential for both data reduction technologies like deduplication and compression to reduce the footprint of that application workload, but then also if you have effective copy data management and you can reduce the footprint of those copies, what do you get? And the answer is you see flash this year becoming the crossover point. I don't mean all flash, or I got to clarify here. I don't mean every flash product because you have to have the right feature set in the array in order to do this. But if you've architected properly, if you've built that right feature set, as we've done with Xtreme IO, this year is the inflection point where you can cross over and have massive savings. And that's only on the cost side we're talking about right now, the savings and productivity and being able to get things done and eliminating very inefficient processes in a company, those are just as strong. Yeah, Callaway was talking about the all flash data center saying they can't wait till it gets here and he actually said we see the day that it's coming. Two years ago, I remember I think somebody from EBA I heard give a talk one time about all flash data and people were laughing, so that's ridiculous. Not so ridiculous anymore. It's not at all. First of all, it's incredible from just the environmentals, right? Space, power, cooling, all the things people are worried about physically in the data center. But for us, what we really see is that it's a huge enabler for customers to do things that have never been possible before. So Callaway is a great example. They use Extreme I.O. in their SAP environment and SAP is a typical enterprise application in the sense that you have SAP production but then you've got landscapes all over the place for everything, many landscapes for development, many for tests, many for analytics, many for reporting and then there's multiple different SAP applications within the SAP suite each with that same set of many, many copies. And you think about what people have to do. If I have a 10 terabyte SAP production instance and I need a copy, I'm literally sitting there and moving all that data within an array to make the copy and then I've got double the capacity consumption. If I have 10 copies, it's 10 times the capacity consumption. That was a massive opportunity for us on the Extreme I.O. platform to figure out a better way to do it. So once you have an array that's doing inline deduplication and you're representing how the customer's view of the information looks through metadata not through just physical, having all those blocks have to be stored multiple times, we can now do very sophisticated things with how we manage that metadata so that copy services in the array mean rather than brute force copying a 10 terabyte production instance, we do it instantly through metadata magic. I want to talk about, go back to think about last year at EMC World, you guys pull off the gloves, a million dollar guarantee which I think David Goulden, maybe it was Jeremy jokingly said we haven't paid yet. No surprise, of course we didn't pay it. It was a lock. But you really sort of went after this notion that architecture matters, you sort of called out some of the competition. So I want to talk about the competition. You got you guys, there's a number of all flash array companies, a lot of the big guys, whether it's IBM, HP, others have their sort of all flash strategy baked out, at least it looks that way. Talk about architecture, what matters, you talk about third platform, why is your system the future system? Sure, when we talk about architecture matters, what we mean is that there's a big difference between listing a feature as available and how it's implemented. So a great example of this is deduplication technology. You're going to see everybody come out and say that they have some kind of deduplication technology in their array. But the question is, how has that been implemented and how does it work under the kind of load that you expect to be putting on today and tomorrow's all flash array architectures? In Extreme I.O., we made decisions very early on in the design of the product that all of our data services are going to be in line all the time. That's part of what that guarantee that you're mentioning David Golden put out for a million dollars. The reason we did that is because the only way that we can achieve scale and consistency and predictability of performance is to know exactly what the array is going to do on a per I.O. basis every time, keep the latency consistent. So deduplication is one of those examples. We have to deduplicate every single block of data coming into the array in real time as it's arriving. We don't batch it up, we don't commit it down to the flash and then go back and try to comb over it later. So technically you can say you have deduplication no matter how you've implemented it, but there's only one way to implement deduplication that's going to scale and that's going to give you that consistent and predictable performance and that's the way we did it. So I got to ask you some customer examples. Give us some real customer examples of how this is rendering out into value. Couple different use cases would be great. Sure, great example. So Callaway is a great one, but you already know that story. So another one we have a company called Bailey Gifford. They're actually here with us talking to customers and other customers at EMC World. Bailey Gifford is an investment management company and they have a production oracle database. They began with Xtreme I.O. with kind of the classic use case. We need to make production run faster and Xtreme I.O. is great at that. But then we started talking to them about what we could do in this arena of copy data management. So we asked them, you know, how many copies of the database do you have? And why do you have that many copies? And the answer always comes back the same no matter who it is. It's, you know, we need the copies for development and test purposes and we don't have more copies because it's too expensive to make them, it's too time consuming, the impact on production while we're making the copies is too high. And we said, what if we could change that for you? Would that have business value to you guys? And they said yes. So today, Bailey Gifford is running 50 distinct copies of their production database, all space efficiently on Xtreme I.O. So it's not that they need 50 times the space to do that. They need a very small amount of incremental space for each one of those copies, but they're all full performance. This changes the whole landscape of how they do their development workflows. Each developer now has the full copy of their production data as a sandbox that they can do development in. They can now test in those development copies with the full workload that they're going to have once it rolls out into production. So they're getting their code written faster and they're moving it into production with fewer bugs because of what we're able to do on the Xtreme I.O. technology. It's never possible before. The business value to them is tremendous. You think about anybody who's developing code, if you can get those releases out faster that has an impact on your business. There's huge developer productivity impact which ripples through the entire organization. So at most expensive people you probably have in your organization. You got these highly paid, highly skilled workers. You want them to be very productive. You don't want them sitting around waiting for things. Okay, Josh, thanks for coming on theCUBE. Congratulations on the phenomenal year. Looking forward to more examples from customers, more sales. We're going to study, we're putting out. I don't know if they've mentioned it to you but we're looking at the impact. Pretty significant, congratulations. Yeah, thank you guys very much. Great to be here and thank you again. Good to see you, Josh. Looking forward to next year. All right, Xtreme I.O. blown it away. The locomotive of innovation and revenue for EMC. This is theCUBE bringing you live coverage here at EMC Rural 2015. We'll be right back after this short break.