 from San Jose in the heart of Silicon Valley. It's theCUBE covering Big Data SV 2016. Welcome back everybody. Jeff Frick here with theCUBE. We're wrapping up day one at Big Data SV in downtown San Jose. Had a great lineup of guests, practitioners, a lot of technology people, some leaders that are kind of leading the way. We had Paxata on that actually launched their company in 2013 on theCUBE, which is fun to get an update from them. But now we're going to kind of have a little round of table, a little wrap, kind of review what we heard, what we think, and so we've got the whole team here. We've got George Gilbert, Wikibon, Big Data, analysts and analytics analysts. Peter Burris, our newest addition, running all the research now for Wikibon and of course everybody's favorite, John Furrier, founder of Silicon Angle and Silicon Angle Media. So first off guys, great day. Terrific lineup of guests. I think there wasn't a lot going on anywhere in San Jose except for here on theCUBE today. They started late across the street. I think they opened the exhibit hall. We pumped out more CUBE gems and CUBE cards and CUBE interviews and all content combined here in Silicon Valley, I think. If you go to CUBE gems, Twitter hashtag CUBE gems or CUBE cards, great interviews to have. I was very impressed with the guests we had this day and we got two more days of this. So again, on the ground, right on the front lines of all the actions, I thought there was some great points being made and a lot of insights coming out of theCUBE. And again, you're seeing a lot of signal coming out of this market where I was skeptical two years ago. I was really much looking at this and kind of thinking that there was not a lot happening and that the flowers weren't blooming. Going back to Michael Olson's 2011 speech when he announced with Ping Li from Excel Partners that they had a $100 million fund for big data apps. Big data apps just never happened. And so going back to 2013, even a couple of years ago, I'm like, this thing might not hunt. This thing is the dogs are not chasing down the right value proposition. So Hadoop might not happen, all the uncertainty but really I'm encouraged because that was just kind of the shifting of the winds and it's clear that the momentum in big data is there. We see IoT and it's finally crossed over and I think that you have a tailwind in this market where big data and analytics and the software is just not about Hadoop. It's about the diversity and it's about the cloud. The cloud is the engine of innovation. All the companies that I talked to that I was impressed the most with were the ones that had either innovative licensing models and or pure cloud deployments. That is going to enable a slew of analytics and I just think that the mature technology pieces coming together is phenomenal and you're going to start to see the past of business value and ultimately the operationalizing of it. Yeah, I think that integration with cloud is really a huge enabler for big data. We had Marcy Campbell on from Kubell and she said basically they're running their application. It's a cloud-only application. It's just on Azure, AWS and the Google Cloud platform. So cloud clearly a big piece of the puzzle. But let's jump in, ask you Peter, why didn't we have the app thing happen and what's been happening instead? Well, I think there's a number of reasons why we didn't have the app thing and I think it's also leads into the observation I have about what was very interesting about today. And one of the reasons why the app thing didn't happen is because to build an application you have to at least historically know something about the underlying process. And so it's pretty easy to go from understanding accounting to building an accounting application or understanding HR to building an HR application. It's very different to go from understanding the customer to building a what does what for the customer. And I think that's one of the biggest challenges is that we're application development has historically presumed a data model and then we've built other things around it. And here we're talking about not presuming a data model and rather just loading data in and finding insights. And I think that leads to at least for me one of the big takeaways, which was we heard a common theme from everybody even as we heard multiple different approaches to how to go about do it. And that common theme is we have to make this easier. We really have to find ways to make it easier. A recognition that there is business value that's being generated by this tool set episodically periodically with a lot of heavy lifting but that we need to make it easier to actually use the tools and integrate them together. And I know that's a big theme of your George, but also we have to do a better job of bringing the community into how these tools are used. Make them easier to end for end users, make them less onerous, less dependent upon data science or state data scientists and the magicians associated with big data. We probably heard that probably the eight or nine guests that we had today. We've probably heard that out of seven of them. Yeah, it's funny you mentioned that a buddy of mine, Martin Crew from Bootstrap Marketing, long time industry vet a couple of years ago and I was like, yeah, I'm going to Hadoop World and I talked to Bill Sparrow, I was like, what's Hadoop? He's like, he rolls his eyes. He's like, there just aren't that many Hadoop people to implement for the buzz. So I wanna ask you George, you spend your days buried in this stuff. What's your takeaway from today? What are some of the surprises that you heard that you didn't expect? Well, my favorite analogy is Noah shepherded all the animals onto the Ark two by two. And basically with Hadoop, we also have a zoo. They come in threes because of high availability and we also have three zookeepers to keep them sort of shepherd all the animals in the Hadoop zoo together. We're beginning to hear people solve some of that complexity. Sometimes it's in the cloud and they can shield some of that from the admin side and we see some on the tool side addressing the data complexity. But I think everyone understands that this is the first pattern that we have to master in terms of getting barring from Peter. We have to sort of learn to extract questions that we didn't know about. And once those are repeatable, those, when they're repeatable, we can start building applications around them. The ones where we can take customer interactions and we can anticipate and influence them. And that takes the data lake and rather than the ad hoc activities on a very complex foundation, that makes the data lake more of a production platform for repeatable insights in the form of predictive models. And I think when we get here next year, we'll probably be a good deal closer to that pattern where we'll start hearing vendors and customers saying that's within sight. Yeah, and then Shod gave us 250 billion reasons why that's important too. I want to go back to you, John, really talking about the business value. I think it's kind of interesting. When people talk about Hadoop, it's always Cloudera and Hortonworks and Jack Norris would be remiss if we didn't mention Matt Barb. He's just busy out selling customers and doing this thing. But we had Series A funded companies today, Series B funded companies today and Series C funded companies today. So clearly a broad range of innovation, kind of bucking the trend where money's a little bit harder to find. How do you see kind of the opportunity for new companies to drive innovation and find new places to deliver business value? This is the holy grail right now in the market. And the theme that came up was new and Peter has been teasing it out on the intro. What came out of a couple of interviews was the valuation of the data. That came out to be a big deal. And I bring that up in context to your question because the data now is super valuable and there's no real methodology in practices, some of the other stuff that Wikibon's putting some research around is, how do you value the data and practice to business value? And so I think that the opportunity for startups and companies is gonna take a long tail distribution, meaning in the old days, back in the client server days, with Oracle and these big companies, there was a renaissance of application development. And those application develops were full stack, full siloed out companies that went public, people saw it when the list goes on and on, then they get consolidated away. I think you're gonna see a similar thing now where it's a thousand flowers bloom, we're gonna see, I believe we'll see an application renaissance, probably not the way that Mike Olsen thought it would be, but I think you're gonna see pre-packaged unique domain expertise applications that sit on someone else's cloud or stack. That provides value. So you're gonna see a little boutique, like profitable companies, cash flow type companies, and then you're gonna see people that can really add significant value and pick the use case. We had Interana on, which was very impressive, their large scale pattern recognition, that is a platform like business, not just a tool or using analytics. So again, you're gonna have a range of long tail from really big, powerful, popular, highly valued companies down into a niche market of apps that do a great specific thing with data. And I think fundamentally it comes back down to the value of the data, multiple data sources, and then some sort of power source, aka cloud. So I think you're seeing now that maturity take place, and that's the path I see enterprises wanting to take and trying to figure out how to get there, whether you call DevOps, mobile first, whatever, and then ultimately the cloud powers it. So Peter, a lot of things that you've been talking about in your research, I see happening. I think it's a great opportunity, and if you don't have value, you don't have customers, you don't have revenue, and that's why the market today is confused on who's worth what, because they can't value the data, and if someone's not thrown off cash, they will be either accu-hired or go out of business. Peter, I wanna go back to you. I think you actually had the comment of the day today. At least it really struck me, and I hadn't heard it before, which is really comparing data scientists to chauffeurs or operators. That's not the right way to show for us, et cetera. Not the way to really think about the problem. That's fine, and I like that, but I wanna go down a different path with you right now. When we had Amit Walia on from Informatica, he went through a laundry list of historical databases and historical infrastructure really in the context of things continue to change, but there's a high-level theme that you've mentioned before, and I want you to dig into it, which is really, at its most fundamental, schema at read versus schema at write. How is the ability to flip that model impacting what people can do with this technology? Oh, it's had enormous impacts, and I talked a little bit about the idea of the degree to which you know something about the schema in advance versus not versus the question in advance. So is the model uncertain or certain, and is the set of questions certain or uncertain? There's always been a need to be able to address problems that didn't lend themselves automatically to a data-model-like structure, but we didn't have the technology to do it. And so the way it's flipping is a couple of things. I think we can start off by the conversation we've been having about developers. First, the historical model for development started with the idea of understanding your data, whether it was back in the days of Kobal or more recently Java, you end up taking a look at the data, modeling the data, understanding the data, and use that as a basis for then creating a persistent data store that the application went off against. What we're talking about here is something slightly different. We're talking about having the data be available and then asking the developers to find ways of creating value out of that data without necessarily being able to start with a data model to begin with, but yet not being wholly reliant on a data scientist to do magic. We heard a number of the Percata conversation, for example. We're going to hear more tomorrow about this. There's a whole bunch of folks coming on tomorrow that are talking about how data is going to flow across the data plane in an organization in ways that make it easier to create value out of these technologies. But it's a pretty big challenge right now and that's one of the things that's going to happen. We're going to discover over the next few years what are the problems and see the technology in the tool set fill in behind those understandings so that we end up with a real set of platform technologies that support new classes of application development. Peter, thanks for laying it up because I was just going to ask George one of these fundamental themes. And again, you watch this all the time you're looking way down the road. You saw this, which was Spark. And again, Spark's another technology. There's always another technology, but fundamentally, it's really about data at rest versus data in motion. And now combining them. But I want to key off something Peter said about, it's harder to build apps now because all we have is a mess of data. We don't know the structure of the data where we say, okay, we're going to take an order and then credit check. That was a very structured way of building things. But where we probably are going to see the first apps and I'll come back to Spark is where we have like a little mini sort of domain where it might be tell me, I have this data about how telco customers behave. Help me predict when one is likely to churn or help me predict fraud with a credit card check. And it's not that you're going to see big companies sort of trying to sell huge apps. It's more like someone, like a service centric mini app provider will sell these solutions that plug into existing systems of record. And I think we'll see those as the first. And it's not clear that those will even go across industry. Like the churn app might be, there might be one for telco, there might be another one for credit cards and recommendation systems. You might have one that's good for an e-commerce site that is apparel and you might have something completely different for books. Not clear that that's the best example but we won't see these big horizontal ERP type applications around predictive analytics anytime soon, I don't think. Well, Chris from Data Robot, they have what, first interview of the day, I think six million algorithms they said and it's, which is crazy. So to your point, very highly customized for specific applications. Well, that Data Robot one was interesting. One, bots are hot in the world in the DevOps because bots can automate things. That's a DevOps concept of orchestration. But they bring a different perspective where they're automating some of the data science things. And what I liked about this, that interview was this notion of patterns at scale and looking at that notion and the comment I liked from him was, math is vertically agnostic, meaning the whole science behind big data doesn't really look at anything other than computing the data. So that brings back the question of the quality of the data and so you have these opportunities to use the algorithm, machine learning, some of the math behind the science of automation, the future of AI and all this cognitive discussion, certainly IBM talks about, is the math. So that's vertical, not so it's a huge opportunity, but if you applied wrong, you're gonna get wrong outcomes. But John, I wanted to ask you, I wanted to take you down a different path because we see lots of innovation in the technology but probably more exciting for our audience is innovation around the business models. When we talk about, Ryan said that the Caesar's data was worth a billion dollars, unfortunately they didn't know until after they went bankrupt. There's a lot of interesting innovation in the story, I love the cell phone connecting to the smart cities so they can manage the traffic patterns which I'm so tired of hearing about people turning their cell phone. But there's innovation on the business models too. So what are some of your thoughts as we've explored that with some of the guests? Yeah, I mean I think this is one of the areas that certainly is not unpacked as much as it should be in my opinion. I think the business model innovation is just as historic as some of the technology enablement that we're seeing in the maturation of the tech. The business model things is interesting and here's why I like it. We are living in an era of doing something in a new way. Everything that I like and that gets my attention is companies that are doing things in a new way, not the old way or if they're doing it in the old way they're coexisting in the old way like these systems or what not but they're doing something new whether it's looking at the progression of digital was the interest on I love that notion they have huge clients, they have Tinder, Reddit, these big web scale coming to billions and billions of things going on a day and looking at patterns. That involves real time streaming that requires a lot of tech but the fact that it's the business model that's driving their innovation, not the tech, it's the reverse of what most people will think. So you're gonna see innovations around that and that's gonna spawn a new class of business models in my opinion beyond what we see with cloud which is subscription based SAS. There's a ton of documentation out there on SAS business models, certainly that's relevant but I think you're gonna start to see a new transactional business model and this is something that Peter and I have been talking about Dave Vellante around looking at how to put research around that because it is not a siloed vertical app. It's really the vertical app is the application but there's a DevOps, there's orchestration, there's a lot of stuff going on to make that new way monetizable. That's the business value. So to me I think the business model innovation in and of itself is a massive conversation is a great, so huge concept. Peter, your thoughts on that because I mean we were just talking about that. Yeah, it's interesting. There's a small piece about the business model of taking open source software and finding new ways of creating value with it so that folks are willing to pay for some set of services but there's even the bigger question of business model and one of the things that I find absolutely fascinating is that for the first 10, 15 years of this web-based businesses, most of the business models were predicated in advertising and it was almost like everybody was acknowledging well we're really not delivering any real value here and so people will use it but they're not gonna pay for it because nobody will pay for it as big data starts to turn the crank and as we learn how to utilize the crowd as we find new ways of adding data so that the services are more pointed, more specific, more personalized we may start to see a flipping of that and also as we recognize as individuals that privacy matters and that the data that we are giving up as we use some of these free services might also be something of a burden to us in the long term so watching that big picture data model flips which may be happening, I mean some of the conversations about Apple and the FBI and others it that we may start, we may be seeing this first shift in the idea of we're gonna make all the services are gonna be monetized through advertising versus hey let's find ways to actually create something that's valuable enough so that people want to use it. You know that's a great point Jeff I'm just gonna add to that color to that is my feeling that I've been trying to figure out how to talk about this publicly but you just gave me some ideas about how to talk about it is that if you look at all the past two years every keynote we go to Uber, Airbnb everyone uses that kind of as a poster child of what their future could be and ultimately if you look at it Uber's success is not really translating to the Uber of blank. It's really not, I mean you've seen some use cases of but Uber in and of itself is a unique instance of time between the right architecture, the right team, the right business idea all around a right context. Yeah that's it, it's the context. Uber as a context is extremely well done. And so that was kind of, I wanna say lucky it was just a good timing of people who had a good view on something I think in this big data world that we're living in right now this show, this event, this market there will be an Uber-like company that times the business opportunity that they see with the right architecture, with the right data sets whether it's how they decouple it and I think that company will emerge and I don't think it's out here right now I think I don't see anyone with that explosive combination and I think the opportunity for startups and entrepreneurs is there's an opportunity out there while the tides are shifting that there will be an Uber-like company that will emerge out of this marketplace and I think that's gonna be pretty phenomenal. Yeah let's try and you may be right John there's a counter argument to that because of the overall complexity but I think the key point is as we find ways to simplify that's where the new value's gonna be created and so one of the things to think about is think about for example Uber and this is something that our research is looking at right now whether it's Uber or whether it's Airbnb or something like Facebook we talked about it this morning context really matters what is it that people are going to do together? Facebook, great context stay in touch with your friends has not translated into other business models Uber, great context, get a ride as a social network, drivers, people who want rides coming together has not translated into other businesses maybe it will so context is really, really crucial and being able to capture information about what it means to work with others better than anybody else is one of the central features of big data as we move forward and it's what will or will not facilitate the emergence of some of these new business models I'm gonna cut you off right there Peter we're running low on time and I'm gonna give George as our big data analyst the last word but what I'm curious is if you're gonna get in the arm wrestling match with Brian Grace Lee our cloud analyst because what is leading the charge here? Is it big data? Is it cloud? Can the two be separated? Or it just seems like they're so intimately they are linked and I wouldn't want to arm wrestle them because I'd probably lose but I think the way they're feeding on each other is that the cloud is simplifying a lot of the administrative and development complexity that our previous generation and current generation of big data technologies exhibits and I've said for a long time that Hadoop has to answer the challenge of simplification because the main cloud vendors all three of them Amazon, Azure and Google they are building a set of services that were designed to work together they were designed, built, tested integrated, delivered and operated as a unit or increasingly as a unit and as I keep saying the zoo that comes out of the Apache pen those were not designed all to work together people are bending tools around it but you can see the scenes and you talk to customers who are evaluating cloud vendors they'll say yes, long term that would be the competition what those cloud vendors supply as native services. So we're gonna go all night but we're gonna turn off the lights and the cameras if you'd like to join us we'll be downstairs at the Fairmont having a frosty, malted adult beverage probably so you're welcome to come by or you can come by tomorrow we're in the Gold Room you can come by tomorrow night at our party we kick off at 4.30 reception for about a half hour then Peter's gonna go with his presentation we've got a great Victor panel we've got a great customer panel we'll have some tasty food and drink after that so you're welcome to come by you can also come by during the day we've got refreshments a comfortable place to work and coffee so we're excited these guys could go forever we're gonna go for two more days so save some of that dry powder and thanks for watching I'm Jeff Frick, we are live downtown San Jose, big day to SV it's big day to week you're watching theCUBE it's Dave, didn't come with us this time sorry Dave, we miss you people are looking out for you it's Peter, it's John, it's George we're missing Stu and Brian it's more cloud than you guys thought you gotta have to come next year I'm Jeff Frick, thanks for watching we'll be back tomorrow keep it here on SiliconANGLE.TV