 Live from the Fairmont Hotel in San Jose, California, it's theCUBE at Big Data SV 2015. Okay, welcome back everyone. We are here live at Silicon Valley. This is theCUBE, our flagship program. We go out to the events and extract the civil noise. I'm John Furrier, the founder of Silicon Hangout. I'm Joey Jeff Kelly, chief analyst for Big Data at wikibond.org. And our next guest is Chris Tugood, who's the vice president of product and solutions with Teradata Solutions Marketing. Welcome to theCUBE. Glad to be here, thanks John. Great to see you. So Teradata, we had a great conversation yesterday and again a lot of buzz obviously around this transformation around what Jeff Kelly calls phase two of Big Data. You're starting to see the early adopters, customers saying I want to go faster, run, hurry up and go faster, kind of thing with the industry. So we had a panel last night of VC, talking about whether investments follow the money has been our theme for the week. So I got to ask you, where's the money? And where are you guys doing in this big, going faster theme? There's a theme of going faster. So what's the state of this? That's a great question. I mean, if you really look at this space, right? And really, if you think about it in terms of kind of Big Data 1.0, it's really about, let's capture all the data, right? This whole the three Vs. I mean, we're all sick of that, right? But what we really see this evolution war towards is to really get Big Data very pervasive throughout the enterprise. And we think in order to do that, even today, you've got to get Big Data in the hands of the business users. And there's still this gap. I mean, there's a gap between Big Data. It's still relatively technical, despite we made great advances last year in terms of access and SQL, but you really need to evolve it in terms to a place where it's literally Big Data apps where business users can go in and just say, oh, I have a problem I want to solve. It's self-service. They don't have to go ask a data scientist to be able to figure it out for them. They can self-serve it and get the data. Well, the application space is the one we've been watching or it hasn't really existed for the most part in the Big Data space. I think in 2012, we heard, I think it was Mike Olson from Cloudera saying, this is going to be the year of the Big Data apps. Well, that didn't happen then, but. That was three years ago. Three years ago. So we're still looking for, where are all the Big Data applications? Because I completely agree, that's where a lot of the value is going to be. You're going to get into the hands of business users who can actually do something with the insights. So talk about Terry Day's approach to that. You come from the world where data warehousing, leading to business intelligence and reporting, but now you've made some announcements around Big Data applications. How does that evolve your strategy and move the ball forward in terms of getting Big Data in the hands of business users? Yes, last week, just before the event here, was we announced Big Data apps, really powered by Aster App Center. What Big Data apps are is they're a set of applications that are focused on very specific analytic functions that business users can go execute. Really built, industry specific. So we have it for retail. We have it for finance, for telco, for gaming, for healthcare, and even for travel and hospitality. And simple things like, let's say in the gaming industry, we have a Big Data app called Companion Matcher. And it's as simple as, we look at all of the Big Data, which is all of the behavioral characteristics about how gamers are navigating through a specific game, and we run algorithms within this app to say, okay, how do we match different players together? Because when you have someone you like playing in the game, then they play more often and drive more revenue and usage. But that's just one of them, like in gaming. We have, in healthcare, we have paths to surgery, right? What are the different events that lead up to a big surgery? And how do healthcare providers use those as indicators to say, well, how do I do better service for this individual so they don't actually lead up to surgery? Things like in travel, we have sentiment analysis apps where literally you can get an understanding of, what are the different sentiment components, either positive or negative, for different users based upon their behavior and traveling, either in airlines or trains or whatnot. So we've created all of these apps that really are business user-centric. Literally, it's a UI that the business user can go into. It includes all the application logic, the business logic, all the algorithms underneath, pre-linked and pre-built together, and they select a number of different components. They can run the algorithms and then they get the output directly from there. And they're delivered as really pre-built templates and then our professional services, people can help them configure it for their unique needs. Well, I think that's important because I think one of the things we're seeing has meant the application space take off so much because the idea of out-of-the-box applications, if you have a standard application that everybody's using, including your competitors, there's no differentiation there. So it sounds like your approach is to, these apps will get you some significant value out of the box, but there's that, maybe get you 80% of the way, but then maybe 20% is where you can differentiate by customizing that and bringing some of your core experience to the table and actually adding more value there. Jeff, that's absolutely right. In fact, one of the things that we did in designing all these big data apps was we built out what's called Aster App Center. And this App Center is a common framework or application development environment which has a portal, right? So that's the UI in which developers and business users can interact with. They can search for apps. They can share apps. They can collaborate. But there's a set of services and SDKs for logging and authentication and scheduling and linking to the algorithms. And there's also a REST API. So you can interface it with third-party BI tools. The reason this is important is all of the application logic in all of our big data apps isn't buried in Java source code, right? It's built on top of App Center and so it's very easy to customize and adapt for a customer's deployment and then our professional services folks can help do that configuration. Well, one of the other challenges, of course, as you mentioned, you know, you've got applications that span certain vertical markets. And one of the keys to big data applications is melding the data science with the domain expertise. How did Territate approach that? If you're, you mean, because you're spanning a number of vertical industries. How did you approach that? That's where I suspect some of the customization comes in on the end point with the customer who's really got the domain expertise to come in and customize things. But how did you approach that from, you know, understanding like, for example, the path to surgery. That requires a lot of domain knowledge about the healthcare market and clinical processes. How did Territate approach that? Yeah, so if you look very specifically about paths to surgery, right, there's a number of different data elements that we bring together that represents the holistic view around what are all the different characteristics of how people are navigating through the healthcare system. And then we run algorithms like time series and path and pattern analysis to look for, where are the four or five, you know, preceding elements up to surgery, whether it's a major knee surgery or whether it's a surgery on the foot or major open heart surgery. There are indicators that lead to that. And so that gives insight to the individuals to be able to, you know, tweak and modify. So it has a good outcome for their client. But very specifically how we approach it with the data scientists is Astor App Center is perfect for the data scientists because what the data scientist doesn't wanna do is have the business user coming to them and consistently going, oh, I'm writing the same query, I'm getting the same information out. What they can do is they can discover something and then embed their logic directly into Astor App Center and then even deliver it as another app. In fact, we have a very large telco as well as a large financial institution that have done just that. It's some repeatable apps that they wanted to make sure they got out to all the business users so they embedded their logic and then deployed it as a custom app that they brought to market. So I wanna ask you about some of the battle points that we were seeing in terms of the formation and clear the customers want. They want app stores, they wanna provision easily, stand it up on the cloud, all kinds of consumption trends are happening, which is good, it's good for the customer, it's good for the business, but it's causing a shift. This middle layer pass in the cloud and also in the big data area with open data platform opens up the discussion. The apps are clearly defined, people want apps and the data is important there and then infrastructure came up last night in our panel about converging infrastructure, what's powering all that data? So but what's happening in the middle? Is it a free for all? Is it just going to be open standards? How do you guys see that? Because you're playing in that middle ground where the data has to move around, you got to move compute to the data, all these kind of paradigms are coming together. How does a customer make sense of that? The apps are pretty clear, you got visibility on an app structure, app strategy, converging infrastructure is happening, what's going on in the middle? I think there's a lot of different things, right? You want to be able to provide deployment flexibility for all of the different apps. So whether you're going to deploy it in a on-premises system, which is like a appliance kind of model for those people that want that, whether you're going to deploy it as a software only model where they put it on their own infrastructure and blade servers, or whether you're going to deploy it in the cloud. And you start to get to this model where someone says, hey, I want to do that companion matcher thing, but I just want to run it for a month as an analytic as a service. And so that's where it starts to go. And AstorApp Center gives you that flexibility across all of the different- So diversity of use cases is popping up more and more versus the corner of the can use cases. So you're saying it's like the consumption is creating diversity? I think the consumption is creating diversity, but also the consumption wants flexibility in how they get access to information. You know, another announcement and you brought it up that we made this week was, you know, we're a founding member of the Open Data Platform Initiative. And we're very excited about this. In fact, Scott now from Terade Labs, he's been talking about this for a while. It was like last two or three years. It's like, look, we've got to get this into some standards that really are leveraged across the different vendors and customer communities. I mean, I think this alignment of the Open Data Platform with Apache, you know, to really drive forward the whole idea about Hadoop and, you know, where yarn fits and even the evolution of how, you know, MapReduce continues to bring value in the marketplace. I think it's really important because from a Teradata perspective, we want to see that standardization and not fragmentation. There's lots of fragmentation now because vendors do some different tweaks here and some tweaks there and some tweaks over here, but bringing that together with some standard APIs, the winner is going to be the customer. Because what happens is, you know, it's all about customer choice. So let's chill down on that. So we've been, we've been, it's been great for us. We've been analyzing and strategizing and obviously opining on it. So the naysayers are saying, oh, this is a move by vendors to try to, you know, jockey for land grab. And then others are saying, and we're kind of seeing both sides of the discussion. Okay, the customers want to go faster. So it's not so much a land grab admission that, hey, if we partner, we move the ball faster. So there's clearly those kinds of perspectives now. The people who are the leaders like Cloudera will say, hey, you know, this is, we're good as it is. And others are saying, no, we've got to move faster. So how do you weigh in on that? How do you talk to people when they're not in the know, they're not inside the ropes of the industry who are looking at this and trying to make sense of this big announcement because it's pretty huge. I mean, the names aren't little guys. It's big players. You guys are in there, IBM, CenturyLink, Verizon, Capgemini, but there's some people not in there like Cloudera, MapR and others. You know, I think it's important, first of all, I mean, Cloudera, MapR, Hortonworks, these are all really great partners of Teradata. So from our perspective, this truly is about customer choice and about advancing open source in the marketplace and reducing the overall fragmentation. So I think that customers shouldn't get confused because for me it's not about, oh, do I go with the ODP thing or do I, or is Cloudera something different than ODP? It's about advancing the industry to get more value to the customer. And they ultimately make a choice about, you know, which works better for them. We believe that with the ODP we'll absolutely see innovation go faster because you are able to have this standard. And what's happening is you don't have individuals building different things that are competing, right? So by doing that, you get a common foundation and then we can start to look at how do we build differentiation on top of that? So I think this is great for a consumer. Hold on, just to follow up on that. So I want to drill down on that because we were, I've never been a big fan of the concerns. We've seen that Unix days, you know, it has been kind of a barny deal. We're going to try to do a land grab. How open source, the dynamics have changed significantly. It's all out in the open. So I have to ask you, in your opinion, what do you think, why is it so successful with these models? Because, you know, OpenStack's working, Cloud Foundry's working. So the foundation is it the governance? Is it out in the open? Is it all the growth? Is it all those factors? Because conventional wisdom by computer industry standards are like, well, you know, consortiums, joint ventures don't usually work out. But in this case, they seem to be panning out and this one looks good off the tee, as they say. So what's your take on that? I like it. In fact, if you look at what happened with Unix, right? And that was clearly a consortium kind of driven model. And even us, right? We had AT&T MPRAS and NCR MPRAS. And I gotta tell you, as an organization, we spent a lot of time building value add and working on that and hardening it. And it's like, are we really as a vendor adding significant value on that? No. So as we got to some core standards, happy to have other people drive and work the core operating system. And then what we do is interface with them when we need, you know, special environments. And so for us, it's enabled us to take our elements and add value in other areas versus just working on core infrastructure stuff. And so that's where I see this absolutely involved. And focus is in ground specific use cases. Yeah, absolutely. We'll talk about that differentiation. So how does this impact where Teradata delivers value and how it might change your business? You know, as part of that announcement, you heard Pivotal, for example, open sourcing their Green Plum database. So you're seeing some of the database, the data warehouse go open source in some sense. So where does that, how does that change what, how you look at the market and how you're going to add value, differentiate and drive new lines of revenue for Teradata? Well, if you look at how Teradata is really approaching the whole Hadoop marketplace is it's really about how do we add value on top of that Hadoop structure, right? So we have Hadoop in the cloud. We have a Teradata or Hadoop where customers want to have a package that is delivered ready to run. You know, we've done a lot of acquisitions in this space. You know, we acquired Revelytics and we acquired Hedapt and we acquired Rainstore as an archiving application on top of Hadoop. As an example, you know, one of the announcements we also made was Loom 2.4. And Loom as an example is available for free. You can go to teradata.com backslash, try Loom. You can download and you can go put it into production. So, you know, customers that are having challenges with, you know, metadata and data lineage and data wrangling, they can go take that, they can bring it down and they can start getting value within their overall data links. So Teradata's strategy has always been about how can we look at what are some of the bigger challenges and how do we add value. I think another big challenge is really around how do you shore up that data link, right? So with our acquisition of Think Big, right? We, I know you talked to Rick yesterday, right? So, I mean, it's all about a data lake optimization service, right? And really being able to help people look at what are all the things you need to consider about your data lake? What we see is people are like, oh, I have a Hadoop, I'll call it a data lake. Well, the reality is it's not really data lake, right? You gotta make sure you have all the governance, you're dealing with your ingestion in terms of getting data in there, your security, you're looking at the metadata, and you're really building that out. So all of this, if you look at it, it's like where can we bring value from our experience, right? From Teradata and being in this business for a long time. You know, where can we add services around it to help people shore up and get more value so that they're data lake scale? And so we'll continue to do that. We also continue to build out software that ties together analytical ecosystems like the Teradata query grid. That's all about being able to do seamless integration with Cloudera and with MapR and with Hortonworks and with MongoDB and others to help customers bring together an analytical ecosystem in what we call our unified data architecture. So we'll continue to add value around software, deployment options and services as we're bringing stuff to market. So I gotta ask, because I always, I love the term data lake because I love to hate it, but I always throw the term out data ocean because, and the question of few is a little bit more about the future. I mean, obviously Teradata, you see a lot of stuff, a lot of customer use cases and I'm sure there's dark rooms, you guys talk about the future. Oh my God, the data's coming in so fast. Lake kind of implies pretty big, calm, maybe unless it's Lake Tahoe with six foot waves during the last storm, but ocean is really more of a bigger mass as currents as different properties of the data. So what scenarios do you see coming around the corner that customers should pay attention to that you guys are talking about internally at Teradata that you're saying, hey, you know what, this lake's nice. We can play with the lake and put a boat on there and do some nice stuff, but really it's gonna be a really complex ocean of currents and data and complexity at real time. What are some of the challenges in more dynamic data market that you see that you guys are preparing for or talking about? I think there's a big awakening coming from people. If you look at folks that had deployed Hadoop 1.0 and maybe done it in a way in which they had it in very specific lines of business, it was relatively easy, right? Maybe they're bringing clickstream data into that environment. They had a set of four or five users on it. You know, they knew the refresh cycle. They knew what the data is, but if you really think about this architectural pattern, and frankly, I don't care if you call it a lake or a hub or an ocean or a, you know, I mean, I really don't care. The reality is it's a good architectural pattern for storing data in its original fidelity, but the problem is, is when you suddenly have this ocean, let's use your word, where all of these different sources are coming in and all of these different users are hitting it, if you don't have the governance, if you don't have the metadata, it's almost like these users are gonna have data amnesia. They're gonna have to be looking at the data and saying, well, where did this come from? Who's used it? Who refined it? And so we've got to get that governance and that metadata around it. And that's one of the key roles that Loom plays from Teradata Loom and also that, I think, big group. Do you ever see a metadata consortium kind of emerging where it's a challenging area? No, it's control, it's important, it's critical infrastructure from a data perspective, but is it? You know, it's funny, I actually don't. I mean, metadata itself has been a huge challenge, right? It's not something that, and so you see that vendors have built out some of their own specific metadata. I really like what we've done with Loom because it's about being distributed agnostic. You can run it on MapR, you can run it on CloudArea, you can run it on Hortonworks, and it's available for leverage and execution. As you use that and its integration with things like H-Catalog and the standards, I think you're going to start to see some, but to get metadata across the entire vendor landscape, I don't see a consortium coming anytime soon. Well, that governance component is key because it's got a data quality issue and it relates to the impact of the analytics, garbage in, garbage out, but it's also got the compliance issue where you don't want to run a file of compliance and regulations where you're going to get fined. So it's critically important and a lot of the enterprise practitioners that we've talked to, that's an area where they get stuck, moving from pilot projects, prototypes, they feel like they're in a good spot, and then you get compliance involved and legal involved and they put the brakes on it because they hadn't thought this through ahead of time. And the other challenge, of course, is the concept of doing some of the data warehouse optimization where you're going to use Hadoop to store some of that more historical data, but if you don't think ahead about how I'm going to govern this, what is now we're going to call a data lake, when you go to start doing some more advanced analytics on that data, it's a mess. And trying to figure that out, the time to insight is much, much, much longer and you get a lot of frustration. And I'll tell you, we believe that these core data management principles don't go away. Whether you're doing it on Hadoop in a data lake, whether you're doing it in a Teradata data warehouse, whether you're doing it in an Aster environment. And a lot of those core principles, you need expertise to understand how do you go deploy that? How do you build out the governance on those ingestion rates? How do you make sure you're tagging all of the metadata within the environment? And as an example, we've been doing this with Think Big for like four years and we've been doing it with Teradata for like 30 years. And so that expertise about how you shore that up, regardless of the core technology underneath is really important to be able to get it to scale because what's going to happen, I think people will start and say, well, I have one source, now let me add another and then they'll get to your point, it'll get shut down in bureaucracy and culture and it'll just stay small. For it to really scale, you've got to get all of those components around it. And that's something we can help out with. Yeah, I mean, it's almost beyond, not just moving it into quote unquote production, but moving it to a platform level where you're actually using it as a platform to continually build new applications and new use cases. And for that, the governance is critical. When simple things like immutability in terms of compliance, things like archive. You know how many people are thinking about archive in Hadoop? Not a lot, they're like, oh hey, I got three copies of the data, I'm fine. Well, with our acquisition of Rainstore. I mean, archive where you can start to get 40 to 100x compression. When you're using a system like that, even if it has a lot of that original fidelity data that you don't need all of it for analytics, you're still going to want to archive it and back it up. And those are just basic data management principles. So my final question, you guys are like cleaning up the house if you will. Get that, all that baseline stuff done. Archive is all those concepts that need to scale. So there's like this whole scaling issue. The question is for the next generation, people on the edge of the network. The business analysts, people who are playing with the data, who aren't necessarily the geek, data geeks. They're application users of the data. So what do you see as that trend? What are the key things that people are looking at from a peri-data perspective going out and serving that mark? Obviously visualization and making the data presentable lowest, but what are the key trends that you guys are focused on for bringing it out to the common person at the edge of the network? And we talked about earlier, our focus is all about delivering those big data apps. So literally those different users can solve their specific problem. So if I'm literally like a, I'm in a grocery store and I own dairy and produce and I wanna understand the affinity of different products. You know, it's not like I call back the corporate. I literally have a web UI. I can go into, I can say, I wanna look at all the affinities between dairy products and produce products between this period, run it, and then they don't have to worry about what's the logic or what's the algorithm or what's the format of the data. They just run it, they get a result set back and then they can make a decision. Do I bring these certain products together in terms of floor planning or do I do some specific couponing around it? So for us, it's all about the app because you want to mask all of the complexity underneath it and make it as simple for the business user. And we think the way to do it is by giving them these analytic apps that are able them to answer very specific business questions. That's the app tsunami coming around the corner. I think, you know, I think Michael Olson and Ping Lee, when they mentioned that, they were just a little bit early and a lot of things were going on but I think certainly that's a vector that's happening. So Chris, too good, thank you for coming on theCUBE. Really appreciate it. This is theCUBE, we're live in Silicon Valley here at the Big Data SB event, conjunction with Stratoconference, a dupe world. This is theCUBE. We'll be right back after this short break.