 Live from the Oracle Conference Center in Redwood Shores, California. It's The Cube, at the Next Generation Engineered Systems launch event. Brought to you by headline sponsor, Oracle. Okay, we're back everyone. We are live in Silicon Valley at the Oracle headquarters. This is The Cube, our flagship program. We go out to the events and extract the signal from noise. I'm John Furrier. I'm Mike Coase, Dave Vellante. Our next guest is Neil Mendelson, who's the Vice President of Big Data. Just came back from the speech up on stage. Welcome to The Cube. Thank you. Big Data is big. We've been covering it for a long time. Oracle has, you know, is known in the industry for using data a lot, database, business, database, data. And I'll see you in a relationship with Oracle. It goes back, I think now over three years now, with Mike Olson, the CEO of Cloud Air, was excellent Oracle, sold Sleepy Cat. So you guys are in the ecosystem. So, and your background in data warehousing throughout the industry is known. And we've talked about it off camera. So my first question I want to ask you is, what is the big shift and an inflection point going on in the industry? Because there's a shift and an inflection point, right? So everyone's shifting to a new modern era, but yeah, there's some inflection point that deals with new technology. What's your take on that? Well, I think, you know, what we've seen over the last few years is we've seen this emergence of net new technology, right? And it's emerged in, you know, in the corner, right? It's emerged with a few really, really smart guys in the corner working on this special thing, right? It's different, it's special, it's separate, right? And now what you're seeing is that we're making a transition into the mainstream, right? Where now Hadoop and NoSQL and with the advent of appliances like what Oracle's offering with the X5, we're really able to take that into the enterprise to make it possible to operationalize, right, what was in the laboratory? So a couple of years ago, I think it was two years ago at Oracle Open World, Larry gave a demo and basically he positioned, you remember this, well, he positioned, he was the Twitter data, right? And he positioned sort of, do your filtering in Hadoop and then bring it into Exadata and we'll, you know, power it. And it was a pretty good presentation, as Larry always gives. Is that the general philosophy that you guys have? I mean, your talk today, you sort of seem to be embracing Hadoop into the mainstream. Maybe a little bit more than Larry's talk two years ago, his sort of poo-pooing Hadoop, not poo-pooing it, but just sort of depositioning it as the filter and then doing all the powerful stuff here. Is that sort of the mode of operation today or do you see it differently? Well, we see these things as really being complimented, right, I mean, you know, some of the examples that we were talking about were, you know, in the relational side, we've got, you know, information about customers, right? Transactions and the like, right? And that's only growing, right? On the Hadoop side, we have, you know, new information that's maybe coming in from Twitter and the like, right? And really our opportunity is to combine these things together, right? To be able to seek a new insider, a new business model that we didn't have before, right? So it's not a question of one or the other. It's really a question of combining them together. Well, so you hear a lot in the big data circles, but it used to be batch and then there was spark and, you know, whatever yarn, everybody wants to go real-time. Where do you fit in that real-time nest? So are you the A key ingredient, D key ingredient, a compliment? So one of the things to enable real-time is to, you know, ensure that essentially you don't have any latencies built into your processes, right? So, you know, if you're taking a, you know, a one-platform-centric view, so everything has to be at Hadoop, right? Then you've got all these inherent latencies of having to move information from one part to another part to the third part in order to be able to, in real-time, provide a recommendation, right? What we're seeing is that the worlds of transactions, warehousing and Hadoop and big data collapsing, right? And becoming essentially the fabric that you see across the data center to enable you to be able to take a transaction, be able to provide a real-time response, right? Based on a predictive model that maybe was built in Hadoop. But it's not one where, you know, you're going to be taking months or quarters in order to figure out what you're doing anymore. And you've got to be able to cut down on the amount of moving of data and the number of moving parts. And that's really what we're striving to do. So let's talk economics. Five years ago, we had Jeff Hammabacker on and he told us, he went back to his Facebook days and said my mission was to break this model of the container where I had to pay all this money for the storage container. Yet in your talk today, you talked about economics of, you know, your solutions. You gave a customer case example where it was actually cheaper. So what's happened? What's evolved and changed since those early days or has it changed? Well, you know, Larry really talked about it today, right? Oracle has historically always competed on value, right? And, you know, in the Hadoop world, right? You have to first compete on price and then you can compete on value, right? We've actually been doing that, right? With the big data appliance for a number of years. But it's been a bit of an uphill struggle because when people think about, you know, the lowest possible cost, they haven't historically turned to Oracle. You know the short answer. Oh no, we haven't been, right? And, you know, you hear people actually, you know, are surprised, right? I myself kept going through the numbers again and again, there's got to be something wrong here. Let's do it again. What am I missing? Wait a minute, what aren't you telling me? And the numbers are real, right? Now what we're doing is as a company, right? We're continuing to focus on performance. We're continuing to focus on value. But now we've got a tremendous focus on, you know, driving down the price, the initial price of whether you're talking about a virtual appliance to run an entire data center or a big data appliance. And I think that's going to help us tremendously to raise the awareness that Oracle's in this game, right, in order to compete on price and on value. I mean, there's two strategies you see with the engineer systems, the competing on prices. It's a good competitive strategy. But in emerging markets that are nascent like big data, there's not a lot of value demonstrated yet. So price is important for entry and ingratiating into the market. Okay, get that. So I got to ask you, where is that value now? And we see, you know, people will say the killer app and it needs to be a killer app in some market. But let's talk about that. In big data, what is the killer app? Is it the fabric? Is it the app? Is it analytics? Insight, you really can't put your finger on one killer app in big data or is there a killer app? I think the thing that are focusing on more often than not is the killer is not about looking at what's happened in the past, right? And looking about, you know, in order to understand what's happened historically. The killer at this point is in order to predict, right? It's to look into the future, right? How do I provide that next best offer so somebody picks up an additional product? How do I, you know, rather than using, you know, standard maintenance procedures to, you know, if it's Thursday, I must have to change the filters in, you know, a factory. Instead, we're going to be using this new sensor technology and we're going to be gathering that data in and we're going to use it in order to predict what a failure might occur to get to it before it actually fails. And it's that prediction, right? That people are really looking for that provides that essentially killer. Well, that's a couple of different components but let's go back a little history. In 2011, we were at theCUBE in New York for a dupe world. Mike Olson and Ping Lee of Excel, we're announcing a $100 million venture fund for big data apps. Categorically, it sounds like there's an application market there. It should never materialize. When in reality, the market became, killer app was analytics, baseline analytics. So now you're talking to a whole another level about insight. So this is a fabric issue. It's not so much, this is a big data app. In visualization, Tableau, I can see things like that. Well, we are actually starting to see the emergence of big data apps. So we have, and we're not only seeking them out in order to partner with companies, we're beginning to augment our own applications with big data capabilities. And I'm sure you've seen recently, we've acquired companies that have big data fabrics, right? Like Blue Chi and others like that. So we are beginning to see companies begin to specialize. So in areas, for example of looking at churn prediction and network failures for telco, right? And we're seeing big data apps begin to emerge from companies like Emcentric and others like that to provide an app that specializes in this industry, in that case, telco, on that particular business problem. And that's an evolution. So would you say, I would agree with you by the way, I'm starting to see that as well. Certainly even you guys put the marketing cloud, you're seeing big data, big part of that. And the stacks developing. So do you attribute, and I bring up the reference of 2011 because it's a little bit premature in my opinion, but. Is the stack ready for prime time? Is the cloud ready? So is the sign that we're seeing the emergence of apps a signal that the application market's about to explode or coming on board? Because that means, would assume that underlying technology's ready. Well I think if you go back a few years and if you were an aspiring app vendor, right? And you want to build an app, right? What are your initial barriers? Well one barrier is people say, well what is this based on? Well it's based on this big data Hadoop stuff. People go, what is that, right? So you've got that initial objection to overcome. And then, right, scoop, right? All these funny names, right? This kind of thing. I don't recognize that stuff, right? And then you've got security concerns that people talk about, right? And if you're an application vendor, you need a stable platform in order to build on, right? You don't need one that's moving all over the place all the time, right? And when everyone is out there building their own Hadoop clusters one at a time, each one being completely different. You can't build an application business on top of a moving foundation, right? So you've got to have companies like Oracle and others that are beginning to essentially standardize that big data environment so that app vendors can have a solid foundation to build upon and companies can use that to build their own applications. This is the classic enabling technology platform. So you need a stable platform. You don't want a saying that's shifting all the time and you need an enabling technology. So I've got to ask you, what is the, in your opinion, disruptive enabler for that enablement that you guys are providing? And you compete with some other platforms that are trying to establish themselves as fast as possible. Table stakes is a platform and you need an enabling technology. So what would you guys consider as the disruptive enabler that you're developing? So table stakes is to begin with is cost, right? Because let's come back to the dollar, right? Because everyone's assumption, their working assumption is that, all this stuff is free and it's really, really cheap, right? So, and they think that they can fundamentally build it less expensively than they can buy it, right? And that's the working assumption that exists in the market. And what's interesting about that is that the people that are engaged in this, right, really take, I used to be one of those guys, right? Immense pleasure. Alpha Keys can do it, but they can't do it at scale unless there's a Facebook-like example. Exactly. Not everybody can be Facebook, right? Even Facebook struggles to be Facebook from time to time, right? Facebook has Oracle now, so, you know, but this is the point. Google would love to be the point. The Facebook Google Envy is now off the table because the cost to manage it just on the database side or is that what you, because that's kind of the big issue, right? Yeah, I was speaking to a customer just today, right? And it was talking about how difficult it is to be able to get his own people, right? To think beyond the notion that they can actually create value for the business by assembling the underlayment themselves, right? That he really needs to get them out of the weeds and up to focus on what is going to provide value to the business, because very few companies, if at all, right, can really compete with the likes of ourselves and others, right? That are innovating at that level. Instead, what they need to do is to build off of it, right, is to leverage what we're providing and what others are providing, that stable platform to leverage upon, right? And not focus their IT energy and time, which translates into money, and essentially assembling a jigsaw puzzle. So, Hype aside, would you agree, our research shows that most organizations are struggling to get real return on big data. You got the outliers that get killed, but on average, they're turning 55 cents on every dollar spent, and they're struggling to find their way. And what we're seeing is that those failures occur because they're trying to build stuff themselves, it takes them, you know, they think they can do it really quickly, right? And month after month after month, it takes them six or more months just to stand up that environment and then to essentially build upon it. And once they actually get something, right, wow, how do they actually operationalize it, right? We've seen example after example, where they actually have some real value that they've come upon, but then they take it to the operations group and they say, well, please operationalize this stuff. And they look at this and they go, are you kidding me, right? There's no- I think that's the biggest challenge. Operational integration is huge because big data can always get like, you know, quick anecdotal benefits. When you want to scale it, it's very challenging. And that's really where we come in, right? We come in in really helping businesses, right? Not just get started really quick, but to move from the laboratory into production, right? With the same kind of guidelines that are necessary to operate today. You can't drop all your security and all your- Well, here's the problem that we see. I mean, I see, and I wrote a post on Forbes about this one, about the marketing cloud. And I use that in Oracle as an example and it's going to look more like Amazon than it's going to look like than prefabricated, bloated lights in the software, like the old jives of the world. But because big data hits every department. So it's not like the operational playbooks which are tried and true, and you go to certain companies, here's how we do it in manufacturing, so we do it in HR, I mean, they've been around. But now you have more of a horizontal integration that some companies just can't handle. Do you see that as well? And what can companies do to be prepared to integrate big data and have a data-driven mindset? So what we're suggesting, and you're beginning to see a lot of writings upon this, right? I was reading some things lately from Gartner and others like that, which talk about really the notion that you have an operational side of information technology and then you have an innovation side. The operational side has to be able to take ideas and move them into production quickly with the kind of standards and processes and procedures that you'd expect. But the innovation side needs to operate in a much more agile sense, right? In IT, we've learned to essentially eliminate risk, by essentially covering all bases, right? It's hard to innovate like that. So what we're suggesting is that companies essentially have an innovation laboratory, right? Where they combine their business analysts with data scientists, with data wranglers, from people with the business, where they can bring together disparate datasets and explore and try out things, right? While at the same time they embed big data in their operational infrastructure, so that when insights are derived, they can quickly operationalize this, right? Because if you, you know, one approach or the other, right? One extreme or the other doesn't seem to work, right? You can't have a completely agile infrastructure, that's chaotic, right? Nor can you have a completely structured innovation. That leads to nothing. So it used to be run the business, grow the business, transform the business, and they were sort of three separate investment silos. What you're suggesting is you connect those through insights and actually fuse them in some way, shape, or form. And it's hard for a lot of organizations, you know? In the Valley, we talk about, you know, failing fast, right? And how we learn from failure and that we appreciate failure. No one likes failure, by the way. No, but, you know. In a fail, go do it fast. How many things have you learned from your successes? Me, nothing, right? Only catastrophic failures went, oh my God, I actually learned something, right? Don't do that again. Don't do that again, ooh, that hurts, right? You know, it's only by the contrast that you learn, right? You know, and that kind of mentality has been driven out of us in the IT sector. So we have to bring that back, but we have to contain it, right? So that it doesn't essentially put at risk, right? The information that this is to tell. And the assets. Absolutely. So let's talk about, given that, let's talk about developers, because in order to have that flywheel going, you need to have some activity, which one of them is development. So, you know, a lot of people are developing apps and mobile apps and infrastructure and software to integrate in and build their business, grow the top line. So what is the developer's data mine right now? Because obviously, you're seeing apps come out. That's an operational issue as well as an innovation issue that kind of crosses over between them. So what's your view on that right now? Where are we with developers? I mean, data scientists, I would put it in different category, data wranglers are different categories, but like, developer's going to be for the mainstream users. Well, I think first and foremost, you know, you're seeing far less development happening within individual enterprises today than you have ever before, right? Because people are really opting to buy versus build across the board, right? So, you know, if they can buy it and assemble it and do it more quickly, right, why not do that, right? They're only going to build the things that they actually think that can provide them some kind of a strategic advantage, right? So, you know, you're seeing a growing community of developers outside of businesses, more so than you see it within businesses themselves, right? And, you know, that's kind of the way the world's going right now. My final question, we're getting the hook here and I appreciate you coming on, Neil. It's great to have you on theCUBE. We love talking about big data because it's for an hour. I could sit all day. The reception's outside, get a couple glasses of wine. The relationship with Cloudera is well documented, Michael Olson, we talked about it earlier, but now Intel has a relationship with Cloudera. You have a relationship with Intel. There's a lot of things going on with Intel on you guys and Cloudera. So, what is the relationship with Intel and Cloudera? Can you update us on what, is that important? Is it global? I mean, Cloudera was a startup and they have a huge valuation now, of course helping the Intel, but... So, it happens on different levels, right? At the pure engineering level, right? We've been working with Intel, right, for many years on optimizing, you know, the silicon, right, for the operations that are performed, right? In fact, you know, that's what we've been doing on Exadata platform and others, right? So, that work continues, right? And now with Intel tying up with Cloudera, providing that investment, they, too, are turning their attention into these newer technologies with Hadoop and NoSQL and the like. So, together, we think this forms an interesting triad, right? Where we can essentially work together as a team, right? To push the industry forward, right? And as Larry talked about, begin to use the force of gravity to embed more and more things down at lower, lower levels of the spectrum, right? And that's beginning to happen at an engineering level, right? At a, you know, at a field level, right? Intel's had their own distribution, right? And that distribution was, you know, mainly sold largely out of China, right? And now we're working with Intel and China, right? Together with Cloudera to move those customers that went with the Intel distribution, perhaps on a DIY platform, and to move them over to Cloudera on the big data appliance, right? And bringing that together. That brings a global consumption model to the table. Absolutely. For you, multinational companies. And, you know, we thought that was terrific. All right, well, my final question is, you just came back to Oracle, so tell the story. So you came back, how long you had back now? A little over a year. A little over a year, so, you know, you helped establish the data warehouse business, which, good call, billions of dollars. I mean, it's an industry now. Congratulations. It did pretty well. Why are you back? Okay, and why, do you have the pick of the litter? I'm assuming you can come back, you know, Larry probably was thinking, welcome you back with open arms. Why big data? And why are you back? What brought you back and why big data? Well, so, you know, in the intervening time, I had done a number of startups, right? Mostly in the area of, you know, big data analytics and the cloud, right? And what I saw at the time that I was with Oracle is that, you know, you can do a lot of innovations by yourself, right? But it's hard to really attract enough attention to get enough traction fast enough to really change an industry, right? And Oracle's one of the few companies that actually has the kind of draw that allows that to happen, right? You just saw base, too. Huge, right? In Russell. I mean, customers that were using us for transaction processing turn to us for data warehousing. And we think that same opportunity exists for customers to turn to us for big data. And what better place, right, to go and explore the possibilities of big data than the original data company itself or, right? You wanted to bite off something you could chew on, get your hands around. Yeah. And kind of have some fun with it. Yeah, I mean, this is a tremendous amount of fun, right? Being able to see, you know, businesses transform themselves, right? Seemingly overnight. Who knows where this will end up? No, it's exciting. It's very intoxicating at a very intellectual level. Big data is certainly going to be a fabric, in my opinion. I think it's going to be a huge moneymaker for services for all the ecosystem. And again, we're first generation connected worlds. So, I mean, I think it's only going to get bigger. So this is theCUBE, of course. We're big data. We're all about sharing the data with you, stretching the seeds from the noise. I'm John Furrier with Dave Vellante. We'll be right back after this short break, live here at Oracle's headquarters at Silicon Valley. This is theCUBE.