 Live from New York, it's theCUBE, covering big data NYC 2015. Brought to you by Hortonworks, IBM, EMC, and Pivotal. Now your host, John Furrier at George Gilbert. Okay, welcome back everyone. We are here live in New York State. This is SiliconANGLE's theCUBE, our flagship program. We go out to the event and extract the civil noise. I'm John Furrier, my co-host George Gilbert. Our next guest is CUBE alum, former entrepreneur. Now VP and general manager at Peridata, Justin Boardman. Welcome back to theCUBE. Great to see you again. Great to be back. So I brought that up in the intro about you being an entrepreneur now at the big company because really this kind of also maps the contrast to the Duke ecosystem. This year is to me the seminal moment where we put the line in the sand saying, you know what, guys, we're still going to do some futuristic stuff. We've still got to innovate and invent and innovate, but people are writing checks and it's time to step up and deliver value. If you don't deliver value, you don't get paid, you don't get paid, you'll be out of business. So you can raise all the money you want, pay $2 to buy a sale, which companies have tried to do in the past. Those days are over. So you've kind of been on both sides. You've seen the startup world. You know the hustle, you know the game. Now you're on the big co, and you've got to still hustle, but you have clients to answer to. What is going on? What is the state of the market that you see and then what's Teradata up to give us the update? Yeah, I mean I think you're absolutely right. It's time for Hadoop to grow up and get real and likewise the Hadoop related vendors have to start to build real business models around Hadoop. As it pertains to us, I think one of the things that's so exciting for us is seeing how many existing Teradata customers are also using Hadoop and how can we help kind of foster that collaboration and that ability to kind of work with data where it lives. And some of that data is in Teradata, some of that data is in Hadoop. And a big reason for our investment in Presto is to really produce an enterprise quality, enterprise grade SQL solution for Hadoop. And a lot of that comes from our heritage. It had apt, which was acquired by Teradata, and really trying to make that work within a real enterprise environment. And so that's what we're doing. So the other thing that we've been talking about that's kind of not related to this show in big data, but it intersects with is the cloud, cloud native concept. So the unicorns are out there, as they say, or big data corns, call them here, cloud era is one. But the reality is, is companies like Airbnb, there's companies like Dropbox, you got Netflix, you got Facebook, you got Google, these are web scale companies native in the cloud. So the hybrid of cloud is driving this notion that the cloud native apps will be a big user and customer, if you will, of big data, analytics, machine learning. So with that as context, what has you guys done with your Presto announcement? And how does that relate to the relevance of a customer who wants to write and check, look, I'm going to do mobile, it's got to be real time, I need to asynchronous, I want API integration, I want direct connect internet, I got quality of service, I need to put policy in there, which we heard from Cisco, this is the new normal. What's your take on all that? How do you guys evolve that with Presto? Yeah, I mean, I think it starts with a need to have something that can run at internet scale, which was one of the things that attracted us to Presto in the first place. Obviously Facebook uses it on 300 petabytes of data, they have 10,000 users hitting their Presto environment every month. So it had a great foundation. And then I think the role that we play in all of this is trying to help bring to bear those enterprise features that you mentioned, the things that traditional enterprises need, things like security and governance. And this week we announced a brand new ODBC and JDBC drivers, which is really essential to actually using BI tools against data in Hadoop. And so I think that those are all good examples of where we're trying to mature Presto and provide a real kind of enterprise grade solution that regular companies can use. Not everyone has armies of developers to kind of maintain an open source project, so we want to produce something that is well packaged. So what's the ease of use there? Because when I hear Facebook, Airbnb and these companies, we know they're hackers. They have a lot of expertise and tend to build their own. Not every enterprise can build their own. So they want ease of use. They do. They want the easy button. I want to stand it up very easily. I want agile, I want iteration. Got lifecycle management software. How does that fit into the old school models? Certainly Teraday has been around for many years. How does that all come together? Yeah, I mean that's exactly the value that we think we can add to the ecosystem. We think the gaps happen to be our strengths. So in the case of easy installation, that was the release we did in June. Which gaps are you referring to? So installation is certainly an important one. How do you even evaluate these tools if they're hard to use and hard to deploy? So we built a very simple, easy to use installer. I've used it myself. We allow customers to download from our site a VM sandbox versions of Presto. So they can even run it on their laptop in a self-contained environment and experience it firsthand very quickly. Things like the ODBC driver, which is not a sexy feature of the database, and it's probably taken for granted in most traditional, mature database systems. But in this world, really important that you have a real ODBC driver that allows you to connect BI tools. Which to your earlier point, that's how a business analyst wants to interact with data is through a tool. And Tableau and CLIC and MicroStrategy and so forth are great examples of tools that you can now use with Presto. So Justin, we were talking earlier about how databases take a long time to make an enterprise software to mature, to harden, and to scale. Now we know Facebook has some talented engineers and you're becoming a commercial sponsor. How did you get to this scale so fast? From what we understand, the team at Facebook is pretty small. How'd you make such progress? Yeah, so they have about a dozen engineers working on this at Facebook. We've added close to another 20 to the project. Netflix has also been contributing, Airbnb has been contributing. So there's certainly a community around it and one that's growing. Treasure Data also has been adding software to the project. I think the speed with which we've moved has been, well, initially driven by the Facebook team. They are really, really good developers and they had the benefit of starting from scratch with a very clean architecture. And that clean architecture with really no technical debt has allowed it to make it very easy for us to kind of add to that and expand that software quickly. But I don't want to suggest that Presto's a complete database by any means at this point. There's still SQL coverage to be added. There's still enterprise features like security that have to be added. I think towards the end of 2016, you'll see this start to really crystallize and become a more complete system. I think we're in year four now. I think Kurt Monash talks about, it takes about six years to build a complete database. I think as we get into year five, you'll start to really see that come to fruition. So let's talk about Teradata, a big picture. The market's certainly disrupted. You guys have recognized that the world is moving to lower price points and more cloud native. What's been the customer reaction to that? I mean, you guys under a lot of pressure from customers, what are you hearing? How are you responding? Can you talk about that? Sure. I mean, I think it's different use cases at this point. I think Teradata is a very high performance database system. On performance, on SQL functionality, on just stability and robustness, there's really nothing that can compare to that. And so for real mission critical analytic applications and data warehousing applications, Teradata plays a really important role. But I think what we're seeing is, increasingly customers are also using Hadoop. And so I think that's really just a green field opportunity and one that Teradata really wants to be a part of, hence the presto involvement. And then we also have technology that we call query grid, which I'm sure we've talked about on this show in the past, which allows you to query between Teradata and Hadoop. And so we think that kind of brings it all together and allows you to have a more holistic view of all of your data where it lives. Justin, what's the top conversations you've been having this week at the show? Obviously, this is kind of coming together the ecosystem, 60 year of us being at Duke World, the first year is kind of like kind of the founding fathers prior to theCUBE kind of being founded. We've seen the evolution. What's your take? Because it seems to be this jockeying, we call it the NASCAR race, right? Everyone's kind of jockeying in the pack, waiting for someone to slingshot out. That just never happened. Instead, you've seen the big guys come in, pivoting just as fast, it's not just as fast, but pivoting just like startups. So the benefits of big data in cloud is you can be agile. Yeah, absolutely. It's a little faster for the startups, but not as slow for the big guys. So you're seeing Oracle, you're seeing Cisco, you're seeing IBM, you've got HP, you see Teradata. Is there room to play in the ecosystem in one of the conversations that you're hearing? Yeah, I mean, I think it's an interesting point. I think one thing that becomes increasingly clear to me I think over time is how the Hadoop distribution wars continue to kind of bifurcate and fork, if you will, and go down their separate paths. And each distribution is building kind of their own stack of projects and tools specific to their distribution. So interestingly, I think one of the reasons customers got interested in Hadoop in the first place was it's no vendor lock-in, it's very low switching costs, at least that was the perception. And now you're seeing, I think each of those distributions slowly start to kind of build their motes, if you will. And interestingly, that's been one of the reasons or one of the appeals to Presto is because it's distribution agnostic, it kind of is insulating the customer as these distributions go their separate ways. So it allows them to run on either distribution, they could switch distributions down the road, which has been an interesting observation. We had the VP marketing from HPE Big Data, I'm talking about Vertica, some of the new cases where data is either living in the cloud or coming into the cloud in a streaming form, machine data time series. Are you seeing like a shift from the traditional, let's put a data lake that complements data warehouse and to sort of new use cases where you're working with streams of data and you have to make constant. That's not always batch. Absolutely, I think that is definitely another trend. What are some of those use cases? What do they look like? I mean, I think just your point about having data sort of continuously streaming in, Kafka I think is becoming very popular. It was actually an interesting feature that we learned when we were first speaking with the Facebook team is that they had built a connector for Kafka from Presto so that you could actually query some of that data within Kafka itself. So I think that Kafka and tools like that are becoming increasingly important into that pipeline, into the data pipeline. As a just a quick follow-up, so Kafka is really at the message forwarding, I guess in the form of a log, distributed commit log, just alright a little too techy, but the stream processors like Spark Streaming, like Flink and SAMHSA, they have analytics themselves. How do you see stream processors living next to data warehouses? I mean, I see them as complimentary as well. I think you're going to have a need for that kind of real time ability to analyze that data in a streaming fashion, but I think you're also going to be aggregating a lot of data over time. And that's more your data lake concept or your data warehouse concept. And I think those are use cases that they're just different use cases in my opinion. So I think both of these systems will live together as different steps within that pipeline. Do you see, because right now the streaming data, it's like hot topic, but not a lot of production except the leading edge. How do you see the sort of over the next few years, the relative volume of data between those two types? Fully agree, there's history data and there's fresh data. But how should we think about sort of the relative volume? Well, I think it's still early days for a lot of those platforms that you mentioned. I think they will take time to continue to mature. I think, again, for kind of the business analyst, if we talk about that user for a moment, I think they're still very much have a preference to use a tool, to use something like Tableau or Click or what have you. And I think those are more ideally suited for a more traditional analytic database front end, if you will. So I think they will be still communicating with data in through something like Presto or another SQL and Hadoop solution or a data warehouse itself. And so I think there's still a lot of value there. I would say the weight is still in that direction. But I think the streaming side is certainly supplemental to that. Okay, final question. What's your bumper sticker for this year's Hadoop World Strata Hadoop, big data NYC? For the people writing checks, what's your bumper sticker to them if you had to lay out a tagline for this year? Yeah. You know, I think it's, like you said, in the opener, I think it's time for Hadoop to get real and become an integral part of the enterprise. That's not a catchy bumper sticker phrase. I have to think about it. Maybe get real. It's one of the last bumper sticker we had as a paragraph. Was it? Yeah, okay. But it could be an image, kind of like on Twitter, you know, 140 characters and you put an image. No, but that's good, but basically get real. Yeah, yeah, I think so. All right, well, Justin, thanks so much for coming on. Former entrepreneur now, VP GM at Teradata. Thanks for sharing your insight here on theCUBE. We'll be right back with more here live in New York City, theCUBE, after this short break.