 The Cube at OpenStack Summit at LANA 2014 is brought to you by Brocade. Say goodbye to the status quo and hello to Brocade. And Red Hat. Here are your hosts, John Furrier and Stu Miniman. Okay, welcome back everyone. This is The Cube. We're live in Atlanta for the OpenStack Summit. I'm John Furrier, the founder of SiliconANGLE. I'm joined today with a special guest, John Luke Chatelaine, good friend, industry legend. He's been around many innovation cycles, booms and busts. Now the CTO of Data Direct Networks. John Luke, great to see you. It's your hometown. Yeah, it's my hometown. You have a farmer. As you can tell by my accent. We need to do that in French. Are we going to do that in French this time? We can do some subtitles in English and talk in French. Yeah, we can do that. Yeah, I'm a little rusty. But it's great to see you. So you have a farm out here in Atlanta, hit with a tornado. Literally, the book inside the tornado, we had Jeffrey Moran on theCUBE at service now. Dave Vellante interviewed and wrote a book inside the tornado. You were literally inside the tornado. I was literally inside the tornado. Two of them. And then you crossed the chasm to get to your house because it did the whole, okay. So that aside, great area. Great to see you again. So you've been around the block. You've seen trends come and go. Standards, bodies get built up, taken down. Some things work, some things don't. Open source is rocking. Of course, we love open source. But in the day, business is business and you got to get deals done. What's this show like? What's your take on OpenStack? And obviously it's got traction, good foundation, good vibes. Yeah, it's interesting to see it evolve. I mean, we've been keeping an eye on OpenStack at DDN since 2010, roughly. It was interesting to see how people starting to adopt it. We're kind of sitting by the side of the game field here. We're really not the DevOps shop. So we're just seeing how those trends are moving. What we've been doing is more than watching what the hype says or what people are doing is really asking our customers, what does it mean for you guys? And all traditional customers have not screamed up a stack. But that doesn't mean it's a bad thing. It just means that it's not yet applicable to the type of people we serve. But we don't serve your general purpose data center, I would say. So I've been quite impressed on the product. It's a DevOps show. What do you mean by it's a DevOps show? I mean, DevOps, we know what DevOps is. You say it's not marketed as a DevOps show. You say, and that's just more of a culture? I think it's a culture. To me, this morning, it was like a DevOps grand mass. I don't know who's the Po. But it's mainly after when you hear the words that are being spoken, it's about speed, it's about agility, it's about transformation of the data center. I mean, its speed of deployment of application is being able to quickly change configuration and be able to bring new application to the end user, which is a great thing. It's about evolving the kind of a stuffy, a little dinosaur-ish IT world. But it feels to me that it's more about networking compute today than it's really about storage. Now, this is not to say that it's ignoring storage, but on top of mine, my impression is a majority of people are focused on, OK, how do I deploy so many VMs at the same time transparently and in a very flexible manner so I can adapt my data center. I said it was not applicable to us because yet, because our customers are extreme. As you know, you're not generally going to find us behind an exchange over or behind a traditional application. So the people we serve, the HPC community, whether it's the Argon National Lab or Oak Ridge, they don't come to us for what the stock is about yet, and then the intelligence community and others, media and entertainment, they're paying attention to it, but that's not what they're asking of us. So today. Yeah, and so one of the things we were talking about is some conversations going on in CrowdChat right now from our previous guest, ex-gardener analyst, now working for Red Hat, Alessandro Pirelli, talking about, we need to change the metric. So I would agree, it is a DevOps show. When you see contribution by lines of code as a core metric of success, you know that it's certainly early and DevOps focused. That's what DevOps thing, hey, ship code. And that was last year's model, vote with your code. So I get that. The issue is, that's not how businesses run. They run on SLAs and this throughput is cost of ownership, total cost of ownership, cost per gigabyte if you're in the storage business, you know, real metrics, penetration, market share. I mean, so with all that, how early is this market? I mean, you guys doing big deployments and certain in the storage area. We had David Wright on early from Solid Fire, and he's got an all flash array. Yeah, I think it's still very early. For what I would, again, for what I would, for my past at U of Packard, right? And dealing with the traditional enterprise, I think it's still very early. Good progress very early. I can see some adoption in service providers, right? When you're a service provider, a lot of things can be done behind the scene, right? Behind the curtain as long as it does not affect the end customer, it's important. You can probably reduce the total cost of ownership for the service providers. I'm not sure it's ready to be rolled into a very traditional IT department and be, you know, as readily adopted as the classic type of technology. Little careful about things that claim to be free also, right? I mean, free is not free. And, you know, I'm not hearing a lot about total cost of ownership yet. And, you know, I said that back in December. Yeah, let's talk about the metrics that you laid out. So in 2014 predictions, you had done an interview with SiliconANGLE's writer. What were some of the predictions you had and how are they faring? So I think one of the things I predicted is that. Hold on, I pulled up the list. I shared this earlier. Oh, you have that. Let me find it here. SiliconANGLE. My God, he's going to follow me forever. But I've been right. John Luke, searching it. Here we go. Predictions. Let's first here, I predict something. It better be right though. Like I'm forever gone. All right, here we go. Predictions. Prediction number one. Enterprises will keep drowning in data and starving for information and insights. Oh, it is so true. I think that's going pretty good right now. Looking good off the tee, as they say. Middle of the fairway. You know, it's exactly that. They're still completely drowning in data and have little information. You know, only like 1% of information is really being extracted. Remember, it's early. It's only May. We're not even at the halfway point yet, but so far it looks good. They're drowning in data. They're drowning. They're drowning. But you know, the biggest thing is, you know, pulling that signal out, right? And the drowning is real. I've met people spending from transportation company to a pharmaceutical, I mean, I have a good friend who is a high-level executive in a pharmaceutical company. And he just can't find anything. He can't find signal. And you would think that, you know, those people, they live and breathe by data, right? And they're, you know, they're just still. So the whole, it's very real. I mean, it's very, very real that it's still drowning. And so the magic is, how can you help as quickly as possible find that signal and get it out? But 1% of information being truly exploited is not enough. And that's going to be low latency, real-time analytics. Certainly cloud will help there. That's coming out of the open-stack years that analytics are a great use case for the cloud. Putting stuff out there and then having it, you know, going to come in. Second prediction, number two, enterprises will want to do big data projects while their little data issues are still not resolved. You sound like a psychologist there. Their little data issues are not yet resolved. What do you mean by that? But I mean that, you know, there's too many definitions for big data. I'm not going to go down that route. If we say that big data is kind of machine-generated data plus social-type data and all this, I see a lot of customers saying, I've got a big data problem, I've got to solve it. And you ask the question, I mean, are you just getting, have you taken care of your little data? I mean, your transaction data, which is very well-known, very well-structured. That's all that I was doing. That's all they're doing. That's all they do. He did not agree with that, but we'll... Yeah, they don't. So what I'm... He said, no, we do it all. I know. I don't want to start controversy, but yes, data... He's a Georgia Tech alum, don't go hard on him. CEO from Solifire, Dave Wright. I know, I was a AMRE, so that's okay. What he said is data silos, and then in the future, or in the new world, all data will reside on scale-out database. I'm just asking the question, what do we do with your structure? I asked a question. I asked a question. He said, no, we do it all, but really dissecting his answer, maybe he was not thinking properly, but it's transactional. I mean, that's essentially... Yeah, it's transactional. That's relational databases, that stuff in memory. But even that today, I don't think people are extracting as much as they could out of that very well-structured data. So before you tackle a big data problem, make sure you've taken care of your little data, or at least you've got a plan for it, because otherwise you're going to be chasing this horse and... The little data might be hiding in the big data, too. You know, at the end of the day, both are useful. You remember or talked three years ago, where I told you, transactional data give you context, and unstructured data, non-traduction, it gives you sentiment. And it's both together give you the truth, because context with sentiment is powerful. Next prediction, software defined everything hype will increase in total ignorance of where and when hardware and software-tight coupling makes sense. So Markers and Hopkins was actually pointing this out, and he didn't think it was overhyped based on the quote that the guy from Seth, who ink tank sold to Red Hat, he was like, it's completely BS. So he's agreeing with you. Today on theCUBE, he said that. You know, that was music to my ear. That was music to my ear, because you would think of the guys from Seth being like software defined everything. They are software defined. They are all about software defined, right? So I was waiting for somebody to speak about software defined software anytime soon, right? But hasn't been done yet. But yeah, I mean, the reality is, for the past 10 years, a lot of storage has been software working in conjunction with hardware in order to deliver a given quality of service. Now, where I would knock people on is, people that sell you a right array with a full of ASIC and telling you they're software defined storage. No, they're not, right? I mean, but the moment you've got, you know, thousand line, a hundred thousand line, a million, we've got two million lines of software on our arrays. Right? We're software defined storage if you want. But I'd rather think we're just, you know, the right solution, right? Yeah, then- Well, Mark Hopkinson was saying on our crowd chat, well then, if you think that's BS and you're calling it out on it, and so did the guy from Seth, then all data is big data. All data is big data. Well, big data is all data. Yes. Right. Yeah, you're right. It's like, okay, so Mark, you're right on that one. So in a way, it's, you know, okay, next prediction. Commodity and open source everything pundits will still live in blissful ignorance of a little thing called TCO. Well, that's what I said a minute ago. Total cost of ownership, right? It's not because something I give you is free, right? That running that thing is going to be free. This is what Red Hat's fighting for earlier on. We had the analysts on this. Now it works at Red Hat. This is what IBM and HP all recognize is that total cost of ownership is the game. That is the metric. At the end of the day, everything else is BS. I'm going to write that on the crowd chat. I absolutely think that's totally the case. Whether you call it the shark fin or the tip of the iceberg and underneath that line is the real cost, right? So huge deal. It's hard to really calculate what total cost of ownership is. It's hard because of the self-cost, of course, right? But, you know, it is a game, right? You know, I went back in the days when I was in the email archiving market at large scale, I was telling customers that I could give them the storage. They will still pay to get that storage managed. So I don't think anything has changed. Total cost of ownership is about how much it costs to run something, right? And open source everything and everything is free. Sorry, I've been around the block too long. That's it, that's all I predicted. Now, there's going to be another one. Okay, I'm just writing this down. Overheard on theCUBE. Go to crowdchat.net if you want to participate. I said you were on your Facebook page. Now you're really working. I said you were on your Facebook page. No, I don't post a Facebook anymore. People still use the IE site. Prediction number five. EDW economics will still be poor. Drive ETL for data reduction and inflexible model. Fit hence derive faulty insights. So enterprise data warehouses will not have a good value, what you're saying? Yeah, so what I'm saying here is the traditional EDW, as defined by, I won't name any vendors, but everybody knows what traditional EDWs are, are they storage costs, which is astronomical, right? You know, numbers that would scare you. And that storage cost has driven, which is an economic model, has driven a lot of people to not be able to put all the information that they need inside their EDW to attract the signal. In fact, a lot of time in an EDW, you put subset of data or sampling of data because you can't afford economically speaking to put everything. So I've said it many times to me, extract, transform, and load, which really is really extract, torture, and lose. So, but what's really interesting is that there is an emergence of the BDW, the Business Data Warehouse, and Hadoop is a great component of that. So Hadoop, I've said it many times, is not an analytics platform, but it's an important component in the analytics platform. And what Hadoop is bringing in that space, that revolution, giving a revolution there, is that it allows people to keep all of the raw data that they need in order to then do the analytics work. And that the Business Data Warehouse is going to obfuscate the EDW from a amount of data that they keep and where people do lots of analytics. This is not to say that EDW will disappear because compliance data and referential data may still sit there. Last prediction, the big data universe will have a giant black hole into which security, privacy, and GRC, which stands for global governance, risk management, compliance, have fallen and will stay. Yeah, and I would add security and privacy to that. Security, privacy, and GRC have fallen and will stay. What do you mean have fallen and will stay? What do you mean by that? Meaning I don't see a lot of motion around tools to enable governance to enable good privacy and good security around big data. You and I, I think we had lunch a year ago and we talked about the shadow data. You remember the shadow data, right? And so I'm concerned that in the world of big data, people are going to be capturing feeds from social networks and all this. Got to have information about consumers which agree the consumers are giving freely today but which privacy is not really considered. So there is a big void on tools around helping an enterprise have awareness of the risk and the compliance issue with big data. And none of the tools that I see today are designed to the scale at which they need to be designed in order to handle big data. So it's still a black hole. It's making progress but it's still a black hole. All right, so we're gonna end the same way up on time here but I want to just get your thoughts on industry consolidation and during growth. So we heard that the trough of disillusionment is coming in the cloud from the Gardner guy in terms of open stack and all this stuff. And big data too by the way. And big data too is? Right in the middle of it. It's right in the, yeah, I mean, thank God that everyone's doing all their M&A deals now but so this is going to bring opportunities, right? So Intel made a huge purchase, I mean a huge investment in cloud era purchase. It seems to come out of my mouth just thinking about it. That's a wishful thinking. It is practically a purchase in my mind but it's a huge valuation. It's over the top, it's insane but the betting on the future and the big bet that is in this space around open stack you see consolidation, what's your experience? And you've been on the M&A side, certainly on HP you are a big part of the strategy over there when you're in the CTO office and you're obviously in industry luminary now. What do you see for the buyers and the sellers? Who's buying, who's selling, who needs to buy, who needs to sell? You don't need to name names if you want to name names that's great. No, I don't want to name names because this would be looking crystal balls and I don't shine it every morning but... I'm a big picture though. Who's the buyer? I'm from a big picture, clearly. Is it a buyer's market or seller's market? It's going to be a buyer's market, right? We're probably on the beginning of the trough of the illusion then right smack in the middle for the cloud, my gut feeling but I think the buyers are going to be the 800 pound gorillas that are being threatened, right? I mean I would not want to be sitting on 10 million lines of code to do virtual machine provisioning in a proprietary manner, would you? Because these guys, all the 4700 geeks out there they're going to use things that are very impressed by what the guys from Ubuntu talked about, right? The automated provisioning that they have and all that. The juju. The juju, right? I mean I think this is great stuff, right? And there's another one called Solstack. Yeah. Really, really interesting stuff. He's a great guy, he was on the queue as well. Right, so I think that the big guys are going to keep an eye very, very close on what's happening here and they're going to need to own them, right? Otherwise they're going to get commoditized. What do you think about storage? Obviously pure storage is valuation. I mean, legit or over the top, you think they have any chance to be the next EMC? No. Welcome to the bubble, right? You don't think they have a chance to be the next EMC? Pure storage, not a chance. What's holding, what do you think? How does it kind of play out? Give us your forecast on that. You know, I don't like extremes, right? So all disk is not good, all flash is not good, a little bit of both is interesting. There is no searching as one size fits all, right? There are places where pure SSD solutions are very applicable to accelerate some application. There are places where the economics are not going to work because of the size, the amount of data that you have or the type of profile that you have. I'm really more looking at, you know, I'm a big fan of non-flash solid state, right? So I do think that in term, the next generation memory technology are going to usher a transformation. And the transformation would be that enough solid state memory would be close to the application that the traditional storage would be relegated to the archive stuff. So that's what I see. But it's not going to be an SSD place, it's not going to be a direct to PCIe or whatever the bus of the CPU bus of the time it's going to be, or lots of solid state technology. Whether it's PCMS, DRAM, or VRAM, we can argue that till the cars come home, right? I got my preference, but, you know, Intel as others. But lots of memory, close to CPU, a programmatic model or programming models that move from, you know, far-read and far-write to malloc and free, right? And then storage, taking a backseat as being the archive of all that. Scott Tentz has got a big vision. I mean, I like him, I like his Mojo, I like what they're doing over there. But EMC just bought DSSD, what did you think about that move? It's about people and IP. I'm trying to, I mean. That's not their normal mode move. That's not their normal move, but I think they're probably, you know. Given Xtremio's demand, they need to kind of bolt something in there. Yeah, I think they need patents and they need smart people. Yeah, they certainly got it with that deal. It could also be defensive. Yeah. Right? It feels like they got the cool kids, right? Yeah, it's a great team. I don't know, have you seen the product? I have not. No, I love the presentation. Jeremy Burton did a great job at EMC. Okay, last question I want to ask you is, I'll ask everyone is, why is this point in history in the tech business so exciting? You've, again, been through many innovation cycles, booms and busts and hypes and bubbles and so you've seen a lot. Yeah. Compare this moment in time. Why is it out there? Tell the folks out there in your own words what this is all about. Why is it so exciting? Why are people all jazzed up? Well, I can tell you why I'm jazzed up. I'm jazzed up at value extraction, right? So I think it's a great moment because we are at the moment where information becomes the currency of the enterprise. And we are changing how we make decisions. We're gonna get to a nearer of data and information during decision, right? Powered by big data. So that too may excite me. Ultimately, infrastructure is being simplified and automated. On some aspect, I think we're rebuilding the mainframe sometime, you know? The cloud is all virtual, all automated, partitioned everywhere, virtual partitioned everywhere. Sounds like a giant mainframe to me. I mean, you and I have talked about this years ago. The systems guys are back. The old school dudes like us are back and in vogue. You know, I mean, hell, you know, it's a mainframe. It's like an operating system. The web is now a complete operating. I mean, APIs are subsystems, right? I pulled out my JCL books the other day. I mean, maybe we can put that on resume, you know? I mean, data centers, not just data centers, but now if you look at the cloud, the cloud is a distributed network, right? It's connected. You got some compute everywhere. You got virtualization. I mean, why not just create one big massive system? Yours was funny because we broke the mainframe, right? We tore it apart during the open system time. And then we got all those parts and we put them together and it was good because it brought in standardization and open system and all that. And then we realized that putting it together and keeping it working is really expensive and it's kind of nice when they're all unified. So maybe we're rebuilding the mainframe of this. You know, I'm personally excited. I would agree with you. I think I'm excited because what you said, wealth extraction, I think, or value extraction. I think it's a wealth creation opportunity for entrepreneurs, I think. You look at OpenStack and why I'm not too hard on the OpenStack. I mean, how this thing plays out. It's still baking in the oven, in my opinion. But there is some serious wealth creation, value creation opportunities for customers which would then get rewarded to the companies to create that value for customers. I mean, industries are created this way and I think you're going to see some new industry formations that's going to create a new cast of characters, new players. Yeah. Some old ones will fade away, stay around. But I just think it's a great, if you're an entrepreneur, the chips are on the table. Yeah, and we can do good things with it. That's what I really like about it. I mean, I'm seeing what information enable in underprivileged countries like Africa, for example, where we can bring what you and I consider as normal to the fingertip and some kid in the desert that can now have access to this and the enablement and the all the- The globalization is amazing. The globalization of the use of information for good. I mean, I know there's a lot of information that's not useful, good, but, you know. We're seeing kumbaya here in theCUBE, but I think it's true. I mean, this is all about relationships. Social media certainly is doing that. John Luke, great to see you on theCUBE. As always, great segment to see you, hear your predictions and looking at their on track and your insight into the storage activity and certainly open stack predictions here. This is theCUBE. We're right back with our next guest after this short break. Thank you. Bye.