 We're back, this is Dave Vellante with Jeff Kelly, who's joining me on this segment. Robin Schumacher is here from DataStacks, DataStacks NoSQL player based in Silicon Valley. Welcome back to theCUBE. Thanks for having me. Yes, so, well, give us the update on DataStacks. You guys, obviously, got a lot of momentum, Cassandra has a lot of momentum, NoSQL has a lot of momentum. What's new since we last talked to you guys? Yeah, so Cassandra is now about what, five years old, 2008, so the database itself is five years old. DataStacks has been around for about three years. We drive Cassandra from the open source standpoint and also create and develop and offer commercial products based on Cassandra. Since we last talked, I guess we're up to now about 300 customers that are using our software. We had a very successful D round raise, $45 million that we brought in. We released a couple new versions of our product 3.0, which focused on security, big data security. Our product, DataStacks Enterprise, really has the most comprehensive security feature set of any NoSQL provider right now. And then with 3.1, we rolled in a lot of new developer features with Cassandra 1.2, a lot of ease of use enhancements. So that's a quick update. That's good. I want to come back to that security, but before we do, we were talking off camera about Thomas Carrion's keynote this morning. We heard a lot about Hadoop, Hadoop Connectors, NoSQL, Oracle's Key Value Store, etc., etc. I've often said in theCUBE, Oracle will wait for the trends, all the innovators will announce the trends. Two years later, Oracle will co-op them and act like it invented them. Okay, so as one of those innovators, we've been talking about this for the last three years. What do you make of that? Is that a good thing? Because Oracle's doing a little market development for you. Is that a bad thing? Because the gorilla in the room is going to step on you? What are your thoughts on that? No, not at all. I always found it kind of funny. If you remember, Oracle put out a white paper a couple years back on why you should not trust NoSQL for your data. And literally, months later, they came out with their first version of their NoSQL database. So we think it's a very good thing. I will tell you that in all of our customer engagements, we have not come across them one time yet. Now, the Oracle Relational Database, yes, we have replaced that multiple times. But we've really not come across their NoSQL database once in a POC or anything like that. She's saying, I like how you worded that, you've replaced that multiple times. So take us through an example of when you would replace a conventional Oracle database. Sure. It's a broken record. We see this over and over again with many of our customers, whether they're big or small. So, for example, a company like eBay, eBay replaced a lot of Oracle with us. They have around 200 terabytes now of data running across three data centers of Cassandra data. They see somewhere in the neighborhood of 9 billion writes a day, 5 billion reads a day. They have actually one Cassandra table that's 40 terabytes in size that they access. And the reason that they implement us in so many use cases, whether it's messaging, whether it's showing various customer recommendations, showing like wants and owns on their page, is simply because the relational engine, whether it's Oracle, MySQL, SQL Server doesn't matter, this is not built to scale out to be able to go across multiple geographies to be able to handle the flexible data types that the NoSQL data model can. And so it's not so much again that an RDBMS is bad, it's just not suited for these particular use cases. Yeah, but again, you hear a different message, somewhat different message from Oracle this morning, which is sort of do your filtering into do, you know, and then we'll connect up to it and suck all the real valuable data into Exadata and Exolytics, kind of the million dollar red box approach. You guys don't subscribe to that methodology. It's a very expensive approach for one thing. But yeah, in one of the sessions this morning, Andy Mendelssohn, they said they view big data as the next generation of data warehouse. Evolutionary meaning. Correct. It's the next stage. Whereas we see a fundamental difference. For us, big data is line of business applications. It's not data warehouse. So all the things that I described to you with eBay, that's not data warehouse. That's their line of business applications functioning for their websites, for everything else, to make them money. And so one of the ways that we're seeing this differential happen right now in the marketplace between an RDBMS and NoSQL is an RDBMS is very good for counting money. So we see it in the back end, the counting operations and things like that. But our customers are using NoSQL to make money. It is presenting product recommendations, user behavior, pattern analysis. How can I make money? How can I reach customers more? That's how they're using NoSQL. So the data warehouse business intelligent crowd has always targeted in their vision statements the line of business. And I've said a number of times it was a failed vision. If it weren't for Sarbanes-Oxley, I think Sarbanes-Oxley saved that traditional DWBI business, saves maybe overstating it, but helped it survive through the tough times. It seems like Oracle's view of the world is, okay, now we're finally going to be able to achieve that vision by sort of bolting on Hadoop and NoSQL. You see the world differently. I want to unpack that a bit. Essentially, you're saying the line of business is driving new applications that have remained unfulfilled through the traditional data warehouse and business intelligent apps. Is that right? What kind of apps are we talking about? Take your pick. I mean, I will agree with what Mark Hurd said this morning where your average application is about 20 years old right now. It's not built for mobile and all the other types of things that are occurring out there right now. And so we're seeing a lot of complete app replacements where the app is being completely overhauled or they're beginning with a new app, and they already know that a relational database is not going to cut the mustard. They know that they're going to have to go to a different data management engine to be able to handle the types of things that we're seeing. And those things, again, flexible data model, data coming in very, very, very quickly. Lots of different data types. I have to distribute and disperse the data in multiple locations and be able to read and write to those locations everywhere and have everything synced up. The fact is the relational database is just lousy at that. So of course Oracle, as you mentioned, introduced their own NoSQL database. Let's just take a step back. So how do you view what Oracle is doing? Their whole approach to big data, as Dave mentioned, they kind of pooped big data for a while. Now they're basically talking about how important it is to their future and the future of their customers. My rather cynical take is that Oracle would love to kind of see the open source community who do some of the NoSQL members of the community like yourselves basically just kind of go away. They basically adopted some of the NoSQL technology. They've got their own database. But really, look, their moneymaker is their traditional database. Do you see them making any kind of investment or any kind of innovative use cases of their database, of their NoSQL database? Are they doing anything that really can say, hey, they're taking this somewhat seriously and actually putting a little bit of money behind it? Well, they are. I mean, they're NoSQL database. They are making enhancements. And so they've got a couple of sessions here at the show, and I guarantee you they will be showing enhancements in their NoSQL database. But watch their architectures, their architecture graphics. Watch the way the arrows go. How do the arrows go? They go from NoSQL into, guess what, an Exadata machine, an Exolytics machine. Where am I going to make my money? I'm going to make money on those things. So they're fine for the NoSQL engines to ingest the data and things like that. And they recognize that a relational database is not geared for quickly ingesting the type of data we see all the time. But they also very much want to move that data into their big data appliances. That's the end game, the goal, and that's where the arrows go. And also, you're seeing some enhancements here at the show where they're beginning to introduce SQL extensions that they like to think will obsolete Hadoop. So they've created some SQL pattern matching. They're saying, look at how fast this happens versus what you have to do with Hadoop that's going to take a long time. Right. And so they also, of course, introduced their in-memory database. How do you see that? What is your take on basically the introduction of that? We looked at it this morning, talked a little bit about it. And how is somewhat a reaction or a response to what we're seeing from SAP HANA? It is. But what is your response as, how do you see that type of database both HANA and now Oracle's in-memory database playing against what you guys do? Yeah. There's no question that memory is very, very big. And as a matter of fact, the way Cassandra works today, when you write data in, outside of writing to a commit log to ensure your data is safe, we immediately write into memory. That's where it instantly goes before we ever flush to this. So we have very, very fast write capabilities now. But what you're seeing them tout are the read capabilities. The analytics queries, that's where they're saying, okay, that's where we're getting 100x fold in performance. But notice what they're saying on the transaction side. Maybe a two-fold increase. So they're admitting that a main memory database isn't a cure-all for everything. Well, absolutely. And I think they've improved their messaging from last year. They've tightened it up a little bit. They've got a little bit better story. But of course, it's all about the converge system. And they talked about, we'll take the bottom line. We'll take the investment that you would have to put in engineering these systems in your data center. We'll do that work for you. What is your, what a data stacks approach to that kind of appliance converge systems model? I mean, it doesn't really fit with what we're seeing in the open source community where it's about integrating multiple heterogeneous data sources and tools. But there's something to be said for taking away some of that pain of having to do that kind of work internally. What does your take on that? That's not been our approach to date. We're very much a commodity hardware type of player where you don't need expensive hardware connectors, anything like that to make things happen. And in terms of simplicity, that's one of the beauties of NoSQL and Cassandra. We have a number of our customers such as Constant Contact. They have over 200 nodes of our database running or cluster running. They have no dedicated headcount to it. They don't need to because it's pretty turnkey. You set it and forget it. Well, that's pretty compelling. So let's talk a little bit more about some of your customers, some of your more recent customer wins. You talked about over the summer, Netflix was another Oracle replacement. Yep. Talk a little bit more. You mentioned some of the reasons people do it. What are some of the outcomes then of some of the customers that you've seen have gone in, they replaced Oracle with data stacks. What are they seeing in terms of the return on investment, their ability to build new applications faster, taking into account new types of data sources. What are really some of the outcomes we can point to? Yeah, it's what you described. For example, Constant Contact saved over a million dollars by moving from a very large relational database to us. And they did it in a very quick time, so in a matter of about three months or so. And this is something that we see happen over and over again. The important thing that the customers have to get in their head is the difference in data modeling. The relational database is organized, it's structured, it's data is organized in a very table, row column, very structured format. Whereas in a NoSQL environment, the data model's a lot different. And so to be successful, our customers very first have to wrap their head around, how am I going to model it differently to go from a relational world to a NoSQL world. And once they get that, I won't say they're assured of success, but it comes pretty easily after it. And so again, customers like Open Wave Messaging, eBay, Netflix, Uyalla, they all replaced Oracle with us. And in Uyalla's case, I mean here you've got a company that experiences somewhere in the neighborhood of two billion events a day. I mean, they're tracking every click, every interaction you have with videos on the web, and they need to analyze that data. There's no way, they knew upfront a relational database isn't going to be able to handle it, they tried, it failed, and they've gone to NoSQL and been very happy. So what do you guys got going at the event here? Are you just sort of going under the radar gorilla, you got? You're seeing our presence right here, it's myself this year. We are not a sponsor of the event this year we were last year, but it's just myself attending this event. So we're kind of birds of a feather. We have to come in under the radar, Q-Logic puts us up with their booth, right? We go gorilla too, but that's good. All right, so what's next for you guys? So what, we're going to see you down in New York City at a Duke World. Are you going to be down there? I don't think that's on our event list, I'd have to go check. I'm not sure exactly which events are coming up, but you can't expect from us. We've got a really strong roadmap with the product develop. We're going to have some exciting things coming up before the end of the year. I can tell you that, can't tell you much more other than that right now. You guys had Cassandra Summit earlier this year, right? We had Cassandra Summit, literally, I think we had. I should bring the cube to Cassandra Summit, we could do that. You know, I've got to talk to Billy. I think we had about 1,100 attendees this year, over 800, that's good. Yeah, so actually one thing I did want to touch on that we have in talked about yet is the whole move to the cloud. Oracle, of course, this is another area where Oracle kind of talked about how the cloud was just kind of a made up phenomenon. Now they're basically talking about how they invented it. But from your perspective, where does the cloud fit into what you guys do? In terms of these big data applications you're building, do you see a lot of cloud deployments? And of course we were talking earlier on our editorial segment about some of the potentially chilling effect of the NSA security revelations about people maybe being a little bit less excited about putting their data in Amazon's cloud, for example. How does the cloud fit into your vision, your roadmap? Sure, we have many customers. I don't think it's about, I don't think it's half, but it's close, that run Cassandra and or DataStacks Enterprise in the cloud. And they do so for very good reason. Cassandra is optimized for a cloud environment, masterless architecture, able to go across multiple cloud availability zones, scale linearly, transparently, so you can add capacity online. Our customers take advantage of that. And it pays huge dividends. So for example, last year when AWS had that huge outage in October, Netflix never missed a beat even though one third of their database was unavailable, they were still able to serve their customers because of how Cassandra operates in the cloud. And that's different than taking again an RDBMS, put it in the cloud and saying, oh now I have a cloud database. No, no, you don't. Right, so, well, all right, so we're talking a little bit about Netflix. In terms of, well, maybe not Netflix specifically, but you talked about the applications that you can now run via, thanks to NoSQL. So are these mostly net new applications or are they sometimes migrating applications that they were running on top of Oracle? You mentioned Oracle Replacement, so I'm guessing it's a bit of both. But I'm curious, how do you see that shakedown happen? It is, it's both, it really is both. You've got some customers that do their homework up front, they never go to the relational database to begin with. Others will try to use what they are used to and comfortable with with the relational database. They'll quickly hit a scaling wall and then they'll turn to a NoSQL engine. But we see all different types of use cases. So whether it's customer profile analysis, whether it's time series data that's very, very good, for example, Thompson Reuters that you saw on the keynote this morning, they're a big customer of ours and love data sacks because of how we're able to ingest all the financial data. So many different use cases. Right, so as we were talking off camera, certainly you must have customers that use both Oracle and data stacks. So how do the two work together? Do you do some integration work with Oracle? I mean, imagine any large enterprise is going to have an Oracle database somewhere. So I'm curious how you go about integrating and making the two work together. It's not so much a data movement type of situation where you're moving data constantly from a relational database to us, but more choosing the right database for the right job. So for example, in Netflix case, 95% of their data runs on Cassandra. Data stacks in the cloud. Whereas 5% of their data still exists in a relational database. Might be Oracle, MySQL, I can't remember which one. But again, they're back in accounting systems where the relational database has its strength and makes sense. And so that's not an area you're focused on. You're looking to help customers build new applications that are really or scale existing applications to make use of this new types of data that are available. Correct, all the different types of use cases you're seeing from big data, high velocity data, lots of different data types, lots of data and distributed everywhere, that's where we excel. Lot of action in your world, Robin. So appreciate you coming on theCUBE, sharing with us the innovations around data stacks, congratulations on your progress and we'll be watching. My pleasure, thanks for having me. You're welcome, keep it right there, buddy. So a lot of flash action today, obviously flash and Oracle speeding up Oracle databases is a key theme. Guys from Viren, they're going to be here. Big acquisition, we're talking to Ken Groey. Keep it right there, this is Dave Vellante and Jeff Kelly. We'll be right back.