 Okay, we're back here live at EMC World Exclusive Coverage with SiliconANGLE and Wikibon. This is EMC World in Las Vegas for day two of three days of wall-to-wall coverage. I'm John Furrier, the founder of SiliconANGLE, and I'm David Floyer, co-founder of Wikibon. We are here with Iceland's Nick Kirsch, Vice President and Chief Technology Officer of the Iceland Storage Division. Welcome to theCUBE. Fantastic, glad to be here. So we just heard the president talking about the objectives and just talking about a bunch of things. From how well EMC acquires companies, lets you guys do your thing, they don't meddle too much, you integrate where you see fit, all that's been great. And you guys have been on a rocket ship. Your product has had a great value proposition, customers using it at infrastructure level for big data. And for use cases that are apparently wasn't what it's designed for. You have more big data here, just storage everywhere, and you guys are growing really fast. So first question for you for me is, what are the technical challenges with massive growth within Iceland and an evolving portfolio within EMC? Obviously the transformation message is choice, more commodity hardware to EMC products. What is the technology path for you guys in that picture? Yeah, I'd say the nice thing about Iceland is that we're strategically really already on that sort of commodity storage path. So if you look at an Iceland hardware appliance, for example, it's largely commodity already, we don't use anything like dual ported drives. And so we've been very much focused on enabling hardware commodity components to be usable in the enterprise and adding sort of scale out software features for some time. And you'll see us continue to take that to the next level, which is very exciting for us. So obviously you're very tightly tied in with the Viper announcement and that gives another functionality. What's the sort of thing that you want to bring to Viper, what's the value proposition and how can you help storage users with by integrating in with Viper? Yeah, so the two things that we'd probably provide the Viper layer. One is just a very scalable, easy to manage high performance storage substrate that they can sort of layer on top of. Now Viper is mainly focused on sort of next-gen data services such as Hadoop or object storage. And so there's still a lot of use cases where you might need more traditional NAS protocols. And so the Iceland kind of works in tandem with the Viper to provide those things. The other thing that we're leveraging for Viper is the other part of the commodity vision, which is ultimately Iceland moving to more of a software-only stack. You know, a couple years ago at EMC World, we sort of demonstrated the idea of Iceland in a virtualized environment. And as we make that more real, we enabled Viper technology and the Iceland technology to go in many more places than it currently would go today. So moving it up into the stack, or the majority of the added value into the stack itself, and then what, having commodity hardware as the lowest level? Yeah, I think you can think of it as a two-purpose, sort of taking the Viper data services, integrating them directly into the 1FS operating system, and then taking the 1FS operating system and making it far more portable to many platforms, potentially leveraging virtualization layers or things like open compute platforms so that really that the combination of those two can go anywhere they need to. That's an exciting power. Oh yeah, lots of exciting stuff going on. So what are some of the APIs? What are some of the things that you want to do with Iceland with this new, you've just had a lot of announcements today that's prepped out a little late, I think, this morning. But can you talk about what you've done to enhance the Iceland and get it ready for this new environment? Yeah, I'd say if you look at Iceland, it continues to be sort of a multi-faceted strategy in terms of which customers we pursue. So one of the big announcements here at EMC World was the addition of deduplication to the Iceland operating system, which has been sought after by our enterprise customers. We also announced the availability of file-level auditing and sort of plug-in into partners like Veronis, also very sought after, a bunch of new security features in terms of the ability to allow different administrators access to different portions of the Iceland. You can think of that as sort of the enterprise train that Iceland continues to go towards enabling home directories, archives, these type of workloads. We've also moved more to enable sort of new workloads, sort of cloud workloads. The Viper is one example of that. Another thing that we announced here was actually a built-in object interface into the Iceland platform for Iceland customers who just sort of want to speak object directly and then a RESTful API for control. So you can think of that almost as cloud in a box from an Iceland perspective. So you can be managed by the RESTful API. Exactly, exactly. And that's of course how the Viper team plugs in and manages the Iceland. And then it also enables sort of non Viper installations to work as well. And using whatever orchestration they want. That's right, that's right. Okay, sorry. I was going to say the third leg to sort of the cloud stool is we announced some prototyping around open stack integration. And so we look at the potential for open stack in Iceland. If you look at the storage substrate inside of open stack, it's very immature. I mean, the ability for oneifest to come in and sort of serve out some of those open stack components is pretty exciting. And integrate yourself into, again the same way, integrate yourself into that layout. That's an exciting future. So dig into open stack, I brought up a good point there. Why is open stack attractive to you guys? We're trying to get our hands around that because it's certainly been successful in the buzz. It's attracting a lot of big time enterprises in its messaging. And still early, it's evolving. And Nathan, as Pat Gelsinger said. So how do you guys look at open stack? I mean, what's attractive about open stack to you guys? I think there's a couple things. I mean, first of all, there's been some of these big generational movements. Linux was probably one of those early ones. I was an early tester of Linux. And there was a point where it was clear that Linux was going to be successful. Tremendous amounts of momentum, technical innovation that was starting to exceed or be on par of commercial stacks. You know, I think Android is another example of that that's occurred. And open stack is yet a third example of that. There's tremendous interest in open stack. There's lots of vendor and customer participation to shape that. And the industry really wants sort of a an open source movement to be successful here. And the agendas are pretty clear too. There's not a lot of meddling going on in open stack. No, it's very clear. I mean, everyone knows where their place is. They've all pretty been playing cleanly. Yeah, and I'd say the other big thing about open stack is it's one of those environments that's really been very vendor friendly. And so, hey, there's lots of places for the enterprise vendors to come in, plug in, make the environment more rich, expose their capabilities. And so I think if you compare that to some of the alternative stacks that really haven't gone very far, that's very interesting. It gives us a bunch of- And they're taking the hype out too. There's vendor friendly, but there's contribution mandates. Not implied mandates, not mandates per se, but you know, typical open source. Are you guys shipping code into open stack? I mean, Pat Gelsinger talked about it with Cloud Foundry there, contributing code. Yeah, EMC as a whole now has sender plug-ins for VNX, for VMAX. We'll see a sender plug-in for the Isilon stack later on this year. We're also looking at some potential Swift modifications to make the object layer of open stack more aware of technologies like ours, highly reliable technology, so that you don't have to pay so much of a replication overhead like you do with open stack today. What's the exciting thing about open stack? It's an opportunity for you guys. You guys can get in early in open stack and get in and influence some of the technical issues and help the developers. You can have a nice way to kind of clean sheet of paper if you will, in this future build out. What attracts you about open stack in terms of that regard? What are you looking at to contribute where you can say, hey, if we can do this, rising tide floats all boats, isn't developed, but basically it's a developer environment. Yeah, absolutely. So I think if you look at what we did with Hadoop a year or two ago, where we basically took the Hadoop file system, exposed that of Isilon natively, we made it easy for enterprise customers who wanted to combine sort of file benefits with Hadoop benefits just to take advantage of it natively with Isilon. Now we were early, that market is still developing, but it's the same thing with open stack. I was actually in a customer meeting about an hour ago where they came in, they said all of our developers are now going to be writing to Swift first. But we don't want to go to commodity open stack, we don't trust it to store our data, it's a very immature stack. We would love it if you would just speak Swift directly out of the Isilon. Wow, awesome. That is what you're doing with the next. Which is what we were, yeah, so I actually, I have the open stack developer here at the conference, I brought him into the meeting, we sit here through our prototype, looks like it was a very exciting conversation. But we were both at open stack and we were very impressed. I was very, very impressed that there was no, there wasn't that sort of two hours of this is the way we're going and this is the future. It was, well, this is what open stack is and here are the people that are using it. Talk to them. Yeah, I was there too, it was a great show. It was a great show. I was glad it was a great show. We had three days of the live Cube there, so we talked to everybody, it was great and there's no hype and there's not a lot of noise. A lot of developers, a lot of vendors doing work, so very positive, very, very positive. So the buzz on Twitter with you guys is that the new release of Isilon's 1FS offers virtually unlimited scale out for Viper. What does that mean? I mean, technically, I mean. Sure, definitely. That's a marketing slogan, obviously. So break that down for us. So when you look at, you know, Isilon today scales to 20 petabytes, 144 nodes, sort of one single file system container. Viper is an abstraction layer on top of that for sort of these Hadoop and object data types that can actually aggregate multiple Isolons together. So now I can choose to have several 20 petabyte Isilon buckets where Viper is essentially acting as an object redirection layer. I'm going to keep these things balanced. I'm going to go to this set of clusters sometime where these set of clusters, et cetera. And so you can think of- They can virtualize across the three Isolons in that example, yeah. And eventually I can go across distances after that. So I can continue to just sort of easily scale out my overall object infrastructure. Using Isilon is that scalable, easy to manage, sort of reliable storage tier, and then Viper as sort of the redirection tier for the objects themselves. David, what do you think about Isilon's chances? I mean, you wrote an article I tweeted out there on look at SiliconANGLE and also my hand in the furrier for the folks out there. David Flo, you just wrote an article, free research note, as always, always free on Wikibon. What's your take on Isilon? My take on it was that the announcements were very much in line with Viper. They were symbiotic, that you would, you obviously need a lot of work to do to get all of that out and get that proven and get into the marketplace. So I was very positive about that. My concern is can EMC as a company really drive down the high volume, low margin, open source environment and make money in the way that it's traditionally used to. So that to me is, a lot of good words being spoken, but that to me seems to me the crunch point of when you really got to compete with the ODMs in that environment. So is that what your objective is, is it eventually to make your money on that software and then leave the hardware to other people? What's the philosophy here that you're driving towards? To make yourself really cost competitive? Yeah, I would look at it as really a two part strategy. I mean, there's always going to be some customers who simply are going to benefit from EMC's supply chain, who can't really get hardware any cheaper than we can provide it to them. That being said, our goal at Isilon has always been to take advantage of the absolute cheapest hardware that we can. So we're the only, as I mentioned earlier, the only real enterprise storage system that uses single port SATA drives, so enterprise quality SATA drives, our target for media is to get down to commodity class drives. And I mean the commodity class drives that the service providers like an Amazon or a Google are purchasing and deploying, we want to use those same drives and make them reliable for customers. I mean, that's ideally with the open stack, Swift's of the world are trying to do. We've got years of sort of understanding what it takes to make these sort of media technologies reliable. So on one hand, we're going to make the industry's cheapest components absolutely reliable for enterprise customers, and that's going to drive down the dollar per gigabyte pricing of our normal appliances. On the other hand, we're also going to take Isilon, we're going to virtualize that software layer, we're going to make it available at some point into an open compute environment for those customers who say, look, this is the way I want to go. I want to build my own sort of specialized data center, I want to use technologies like open compute, but there's no way I'm going to get the same kind of software from the open source world. I'm going to buy Isilon for that. And so I really think there's a reasonable bifurcation of those strategies in that the software enhanced, ultimately we're at the end of the day, Isilon is a software company that rides primarily on commodity hardware. We're going to continue to drive that commodity hardware line down, and we're going to give those software benefits directly to customers. It seems to me that one of your potential values is in the hybrid marketplace, where personally I see, I mean data's very heavy, isn't it? You want to keep it together if you can. So if you're in the single data center with the private cloud in that data center, behind their own firewalls, et cetera, and then the more public clouds there, the cloud providers there, if they're using the same infrastructure, the same layer, that seems to me a good way of setting yourself up with the enterprise customers. Is that your long-term strategy? Is that what you're trying to do? You got it. So if you think about why is one of the reasons why OpenStack is so interesting to us, because OpenStack really is the only viable technology that's going to end up in a service provider and on-prem. So the more that Isilon can integrate and make that environment interesting, the better that we can provide that hybrid sort of interaction for customers. We had John Rose on earlier this morning, CTO of EMC, and he's not a storage guy. He's got a background in networking, and when he's talking about the same thing about service provider infrastructure, and enterprises are completely different, consumerization, transforcing everyone to kind of bust out and recalibrate. Now it's not going to be radical, it's going to be incremental. So I want to ask you, my final question is, from your perspective as CTO, Isilon, you guys have been through the battle of these big consumer companies. We know who your clients are, I think, I don't know if we can probably say it, but I think you guys, we know the Googles, the Facebooks, the Apples, the Zingas, all these big hyperscale companies that ramped up, that had needed your product, kind of had to do it on their own. What lessons can you share that you've extracted out of those experiences that you could share with the folks on the enterprise side as the enterprise enters the hyperscale market? What did people break? What did they learn? What were some of the technical tricks in the spirit of giving back? Can you share, and you don't have to say the names, but just best practices of what to do. Certainly, so, yeah, I think one thing that we found is that our technology was surprisingly attractive to this web-scale company. It was at a cost point that perhaps they didn't think that an enterprise vendor could get to and provide a performance they didn't think would come out of a scale at architecture. So, I think that's one thing enterprises have sort of the benefit of. There is actually an enterprise platform out there that's going to be very interesting and affordable for them as they try to sort of get to a new step level and cost. I think the other things a lot of these guys realize is that they have to architect their applications differently as much as they can. They really have to start thinking about the idea that failures do happen in the environment, that we've got to think of a world where I'm going to sort of scale out my application infrastructure. I'm going to take pieces at a time versus sort of scaling out my application environment along with my storage environment. And that's what you see at all the web-scale companies as well, right? They've really taken a, I'm going to build an application that sort of takes a portion of it and I can always have some portion of my infrastructure down, but as a whole, I'm okay because the vast majority of the infrastructure is out. Nick, thanks for coming inside the Cube. You have any more questions, David? That's great. Okay, Nick Kerch, the CTO of Isilon, very big success story. As obviously as a startup growing company and then acquired in with EMC, just a lot of great success in the marketplace, leading the way, a whole nother modern era ahead of you guys, congratulations and good luck with that. This is the Cube, SiliconANGLE and Wikibon's exclusive coverage of EMC World Day 2. We'll be right back with more interviews right after this short break.