 at Big Data SV 2014 is brought to you by headline sponsors, WAN Disco. We make Hadoop invincible and Actian, accelerating Big Data 2.0. Okay, welcome back. We're here live in Silicon Valley for Big Data SV, Silicon Angle and Wikibon's theCUBE, our flagship program where we go out to the events, extract the signal from the noise. I'm John Furrier, the founder of Silicon Angle from Jeff Kelly, analyst at wikibon.org. I'm my co-host and our next guest is Anna Smalltree, director at Treasure Data. Welcome to theCUBE. Thank you so much. Thanks for having me. To Treasure Data, one of the fast growing stuff. We've been watching for a while, but just all of a sudden busted out of nowhere. Just it's almost like they were storing all this data and then boom, straight up in growth and a lot of traction. So first question for you is, where did all this traction come from? Was it just pent up energy? Just kind of waiting to be launched? Tell us a little bit about Treasure Data right now. Yeah, so Treasure Data has a really unique solution in this market. We offer, it's essentially software as a service for big data. So we have technology to help people acquire data and to store it, and then we give different interfaces for analysis. But a lot of our secret sauce is in our acquisition technology. So we focus on data that's created very, very rapidly and we get it into our cloud environment. There was a lot of pent up demand about that. So about that issue of how to stream data into an environment and make it quickly available for analysis. So a lot of our early customers came to us because of this acquisition, the streaming capability, and they also wanted to focus on what they did best. So their core business, analyzing their core business and not necessarily worrying about the data infrastructure so much. So all that, the big bet was on the ingestion side, right, was that pretty much? The ingestion side and the storage side. So we're really good at data management at huge, huge scales. And then rapidly making that data available for analysis. So that people, rather than worrying about building the infrastructure, maintaining it, monitoring, getting the pages, they can just focus on what I call the fun part. So doing all the analytics at the other end where you get value from the data. So kind of, yeah. I mean, I think we've been covering the cloud and big data and we've been talking about, well, when are these two going to really kind of merge? And Trader Data is one of those companies that's actually making it happen. So, you know, first we should point out that you and I used to work together at Tech Target. So we're having a slightly surreal moment here on theCUBE. It's always fun on theCUBE. It's great to watch. It's great to watch. You never know what's going to happen, exactly. So yeah, tell us a little bit about Trader Data in terms of the different ways you kind of deliver your service. Because no, it's not just kind of one size fits all. You guys approach it in a couple different ways. So kind of lay that out for us. Yeah, so we do essentially have one primary service, but people use it in very different ways. So as I mentioned before, we tend to talk about it in three phases. So data acquisition, data storage, and data analysis. Some people are using that all in line. So they're streaming their data into our cloud environment and then they're doing their analysis in the cloud. They're keeping it all in a cloud environment. Other people are putting us as part of a larger data ecosystem. So perhaps using us to ingest all that data that's coming in very rapidly to store it at scale. So billions and billions of rows, sometimes 10 billion rows a day. So that's the kind of scale we're talking about. And they might then do some aggregations to bring down the size of the data and then export the results to another system internally where they might combine it with other data or bring it into another analytics environment or use an on-premise analytics tool. So we have these different models, one where we can fit into a larger data ecosystem and one where it's really the entire end-to-end solution being used in place. And where a company chooses to go might change over time. They might start in one place and then hopefully as they get addicted to treasure data, they're going to start to expand their use of it. But we really try to make it open for integration into existing ecosystems, but also a full end-to-end solution. So if you just want to use that as your primary analytics and data management solution, you can do that too. So providing end-to-end cloud storage is interesting. And certainly as a SaaS, it's a nice business model. We've always been supporting of that. And we love it, it's great scale and it's agile. But I got to ask you, a lot of people have been throwing that under the bus with the whole Nervonix implosion of late. Nervonix was big cloud storage and they went out of business and then literally had two weeks, people that moved their data around. How do you guys address that issue with customers? Well, we love the cloud, but I'm not sure about the data. Is there a product that you guys have there? How do you solve that conversation? So a few different things. We are committed to data portability. So you can export your data from treasure data anytime you want to. We're not trying to lock you in or charge a fee or make it difficult to get your data back. So people will do that on a small scale when they're just doing queries and pushing the results out, maybe to do some other things with that data. People could do it anytime they could take all their data. We hope they don't, but they could. And we're very much committed to being an open platform for that. I think the other main difference is this issue of acquisition and ingestion of the data. So there's a lot of cloud storage platforms where in order to get your data into the platform, you need to use FTP, you need to use bulk imports, and still FedExing disks. It's still a very, very popular way to get data into the cloud. And in fact, I found some news stories about cloud providers announcing the capability to take your mail. And now new capability, we can receive your disk and load it into our cloud environment. So our data streaming technology addresses a very big pain, which means you don't have to stage it and land it and then figure out how to get it into the cloud. But as that data is being created, our treasure agent technology sits right on the servers, right on the web server or the application server and streams that data every few minutes into our environment. We provide reliability, compression, filtering, and transformation of that end node. So when the data gets to the cloud, it's available for analysis within just a few minutes of being created. And it's also, you're not trying to push this huge bulk of data up a small pipe, but you're doing smaller bits of data on a more rapid basis. We've seen great success, but like Splunk, for instance, customers love that product, just allows it to sit there, and then new stuff kind of gets enabled out of that, once they have that functionality. So I want to ask you the same kind of question. For you guys, have you seen some use cases where, because once people get the data and they can act on it, some new insights come out, can you give some examples of where you've seen that happen on your end with your customers? Yeah, so I think people have different ways that they're approaching this. So like I said, there's this way of augmenting your data warehouse with big data capabilities. So we have some customers, one of our customers is a large retailer, they announced a mobile application, and they really wanted to understand sort of how their mobile users were engaging with the product, how that compared with what people were doing online, and how that compared with what they were doing in the stores. Now I can't talk about the exact results because people tend to be a little closed lips about their analytic results, but what they were able to do is get a well-rounded view of the customer. So kind of that old 360 degree view of the customer thing that we've written so many stories about in our time. And indeed. But being able to get this complete view of the customer and understand where the differences and interactions are in a mobile environment versus a web environment versus a retail environment, where some of those connections were and how to leverage that to make more money for their business. So I want to ask kind of a little bigger picture question, kind of playing up with John mentioned about concerns around the cloud. So we've heard about one of the mega vendors in particular was really pushing the big data, I mean cloud marketing, but we haven't seen a lot of it. Is the cloud ready, do you think? Really for to be a place to store all this data, to do a lot of these workloads. And why is that now, do you think? And why is now the right time for a company like Treasure Data? Yeah, so I do think that the cloud is ready. I think there's some types of data which will always remain on-premises and that might be really sensitive data, really regulatory regulated data, healthcare data about my health especially now. But there's some types of data that are going to stay in those environments. But the cloud, since we've been writing about it really, has evolved to be massively scalable, a lot more secure. Even getting data into the cloud is a lot easier than it used to be. So even the increases in sort of network connectivity and bandwidth and being able to get more pikes into the cloud has made it a lot more ready. And you have this explosion of different service providers, new architectures and new platforms which I think make the cloud an ideal place for big data. I think we're also seeing this sort of shift. It used to be a competency to do your data management to store it all in-house. I think the competency is shifting more to what can you do with the data? And now it's a little bit more commoditized to do data management which makes the cloud a perfect place to do it at scale. So they buy millions of servers or however much it takes to store the different platforms and environments and are able to scale that rapidly. And that's really a core competency of a lot of cloud providers that doesn't need to be your core competency as a business anymore. You can focus on the fun part and getting value. I want to ask you a question. You kind of shift gears kind of the strata conference. I know given your background, I'd like you to get to your editorial perspective if we were editing the show. The narrative's coming out of strata. A lot of noise, a lot of signal, a lot of different conversations, a lot of startups that names I've never heard of. Tons. That company never heard of it. That's going to die. That's going to be gone. But let's be critical. Let's edit the narrative. What is the story, in your opinion, coming out of strata this year? And then talk about your story with treasured data. How's that vector into the bigger picture? Yeah, so I've seen the same thing. A lot of names I've never heard. And as tech journalists, we used to have the unenviable or maybe enviable job of trying to figure out exactly what all these people did and really pushing until we understood that. I really feel a lot of emphasis on getting value from data and doing it quickly. So a lot of these tools are focused on either getting Hadoop up and running faster or getting your data ready for analysis faster. So how can we shorten the time it takes to actually start getting value from data? And that's a place where cloud services can really help. Well, there are a lot more tools for doing data analysis on this big data, so being able to process it at scale. So I think we're feeling a lot of emphasis on how quickly you can do it. A lot more tools to enable you in the sort of big data ecosystem. And I think treasured data fits right into that. We're right in the mix there. We get people into production in days, less than two weeks oftentimes. So we're really focused on how soon can we get you to run that first query. And I think I hear that from different angles, there's people who are trying to get you up and running on premises faster or give you different tools to make it easier to use Hadoop. As we've sort of seen that Hadoop adoption curve go up, people have really started to understand and appreciate that it's hard. It's a little bit like building your kit car from scratch. Like, yeah, you can put a bigger engine in it, but you also still have to build the whole thing. So I think we see a lot of interest in how can I make this easier and faster. So here's another question for you. So, I'd love to get you to extract the data from you. So if you have to put people into like crowd spots or like clusters of crowd around certain topics, I see data science is one you see people really chomping on the red meat on that one. What other areas to see the conversations happening around here at this show? In the big data market right now, I see data science is one. What are the big crowd clusters out there talking about? Big conversation pits of interests. So I think that data movement is a big one. So people figuring out how they're gonna move their data around. And that counts just for collecting the data. So the first stage of ingestion, which I talked about a lot, but also for analysis. So are you gonna need to move the data into another BI or analysis tool? Can you analyze it in place? Even things like sharing data. So how you're moving this data around or not. Whether you can leave it in place and do what you need to do with it and share it. So I definitely got a lot of questions about that. I think there's a core of people who are focused on infrastructure and architecture. And how can we process things faster? So getting the performance out of big data processing that is needed. And that depends a lot on how you architect it. So I would say, and then I think the data science and analytics people are very focused on value. How soon can I get value? Can I use languages like SQL that I already know? How quickly can I start to see results and begin to go? I think that was just commenting earlier. Entrepreneur we've been following. Great guy used to work at Bank of America. Tracetic and successful startup. And loves to talk about some of the trends. But one of the things he meant to comment was if you're a database tooling company you might as well just fold 10 right now. Cause that's beyond where we're at now. More people that want solutions and outcomes. Would you agree with that? And what other areas would you say are, hey, old ideas not translating it to this modern era? Old idea, okay. So I mean, I do think we're seeing a shift to more of a build versus buy decision in the big data space. Partly because of the early experiences of early adopters who felt a lot of pain when some of those early Hadoop implementations. So, let's see, can you rephrase your question? So the database he's come as, the database guy should just not focus just on selling the database tools which is an old business model with hundreds of millions of dollars as vendors out there. We know they are. So the question is what old methods are being kind of put out there today as viable ventures? When in reality the market's way beyond that could be point solutions, white spaces. Like what you guys are doing, I think you guys are going to the cutting edge SaaS with storage, very interesting, you got ingestion. Yeah, the old enterprise data warehouse model, if you will. Yeah. The we've written about it. What are the dying business models that are being displaced? Maybe I'll ask the question. What are the dying business models being displaced by disruptive technologies? Well, so I do think that this issue of building it yourself is a dying business model. You don't have to have your whole own infrastructure. You can build things within the cloud. I don't know that database tooling is going away completely. There's a lot of legacy infrastructures that people still need access to. We talked to tons of people who are still on mainframe technology. So I think that some of those ways of thinking about a monolithic of mainframe or a monolithic data warehouse is going away. So we have these much more distributed environments of tools doing what they do best. But it has to fit in with legacy architectures. You can leverage things like the cloud to add big data capabilities quickly. But it needs to fit into this whole ecosystem. So I would say some things are becoming more preferable than in the past, but we still need to integrate with the past. Well, yeah, I think, you know, one of the things we've been talking about the last few days is I think this year we're going to see fewer announcements around new products and a lot more announcements around partnerships and integration with legacy systems. And, you know, just taking a big data space. There's so many moving parts from the integration to the analysis to the output. And then we've got legacy systems, as you say, that aren't going away. Mainframes are still around. You know, there's companies like the Sync Store, which is kind of, it was a mainframe company that's reinventing themselves. And they've been working on kind of offloading some of that mainframe data to big data and Hadoop, whatever. But talk a little bit about treasuredata as I guess your partnership strategy. And a little bit how you do that integration, whether it's from the legacy architecture that people have in their data centers. And then, you know, with the end output, whether it's somebody like Tableau or somebody other BI players. So the first integration point is around ingestion. And that's how we're getting data from the places where it's being created into our system. We have some very unique data acquisition technology that I mentioned earlier. So that's sort of a point of integration. But that acquisition technology is key because it can run on a lot of different kinds of servers. There's actually many different targets that I can output to treasuredata being one of them. So that's one integration point. I think another one is our ability to take data in multiple forms of streaming data or bulk import. Another integration point is this ability to export to your existing environment so that you can, as I mentioned earlier, the case of landing a lot of big data, doing some aggregations to bring it down in size or just get at what you want, and then you can export that to your current environment. And then a big one for us is that we integrate with the front end analytics tools. So our core competency isn't taking in that data, data management, and then making it available for analysis. But we partner with people like Tableau. We recently announced with them a couple weeks ago a formal partnership. So Tableau does the visualizations in the front end. So they are able to reach into treasuredata, either Tableau in the cloud, working with treasuredata in the cloud, or even Tableau on premises, working with treasuredata in the cloud. So this allows people to augment their Tableau environments with new big data capabilities really quickly. They're not the only BI provider we work with. We also work with Metric Insights, specifically on our gaming solution, and that's for digital gaming analytics. This is a really sort of big area. And then we have several other BI providers. We can really support any BI tool. We have ODBC and JDBC drivers. And as our customers request them, which is more often happenings that pushes out into the greater business and analysts hear about treasuredata and say, I could have big data capabilities in two weeks. Yes, I want that. But if we can take their existing tool, their Tableau, whatever interface they're used to working with, and just give them new data capabilities, then they've got an environment they're familiar with. Now they have more data they can work with. Golden, they have a lot more that they can do. So I think we're just about out of time, but I want to ask, so what's going to happen this year with Treasuredata? What's kind of on your roadmap to the extent that you can share? As John pointed out, you guys kind of just hit the scene really in a big way recently. So I think people are just like, where are you guys going with this? What's on your roadmap, and what are some of your big priorities this year? So we've had incredible momentum with our service. Like I said, it's software is a service for big data. It's a managed service environment. So that just sort of as a standalone solution has been extremely popular. And that's why in the last year we quickly passed one trillion rows of data loaded into our service, two trillion and three trillion. So we've had this incredible momentum. We have over a hundred customers now and that's our service being launched a little bit over a year. So the last year has been incredible momentum with our basic Treasuredata service. This year we'll be announcing new solutions. So these are things where we've configured the data collection and we've configured dashboards working with our partners on the other end to provide a complete solution in certain vertical industries. And the two you'll hear about next from us are digital gaming as I talked with. They have so much data. They're collecting data about every interaction a player does in the game. Every move, every purchase when they abandon. So they have tons of data. So digital gaming is a big solution we'll announce. Also advertising technology. So all those clicks and impressions and where we put things and which creative to use you'll hear more from us around that. And there's a few other industries that we have our toes in. I do think increasingly internet of things is what you'll associate with Treasuredata. So we have some early pilots working with sensors so much is instrumented now. And there's sensors in a lot of places that aren't even turned on for lack of a better word. So they're there, they're ready to collect data but people haven't had a place or the wherewithal to actually collect them. So internet of things is where you're really gonna see us in the water. We're definitely gonna be watching that. We've done some research with GE and their industrial internet movement. So that'll be really interesting to watch. And I think it's a really smart move to focus on verticals and solutions. That's what is done. So that's what people want I think. And especially as we kind of move into the early mainstream adopters and big data, they don't want to put together their own to do clusters and do a lot of this integration work. They want solutions. And that's where, yeah, and that's what you talk to you can talk to a business person who signs the check and that's good for the market and the industry, so. Yes. All right, well thanks so much for coming on theCUBE. We really appreciate it, Treasuredata here live inside theCUBE in Silicon Valley. Big data SV hashtag, big data SV. And we'll be right back with our next guest after this short break.