 It's The Cube. Here is your host, Jeff Crick. Hi, Jeff Crick here with The Cube. We're on the ground at the western St. Francis in downtown San Francisco at HBaseCon 2015. We actually came to HBaseCon a couple years back and we wanted to come by again, get an update on what's going on, and we're excited to be here. And we've got our newest guest analyst from the Wikibon, George Gilbert. He's going to be joining us for this interview. Good to see you, George. Good to be here, Jeff. And we have our special guest, John Leach, from Splice Machine, whose CEO, Monty Zwibin, has a very colorful background. Monty's a Cubulum as well. Monty's been on. Yeah, totally. Monty's been many places. But Splice Machine, tell us about your mission and then let's figure out, you know, how that's disrupting, you know, what's going on in not only HBase Land, but traditional database land. Yeah, and similar to what we've been hearing today, it's really about empowering a whole new generation of applications. And those applications, they need the capabilities of relational database to work. But what you've heard, especially in this conference that we've been going to, is people want to scale out architectures. They're starting to transition to the cloud. And they're looking at databases in a different way. And they want these databases to be elastic. They want them to be able to run on commodity hardware. They're expecting different price points because of the scale of the data. And that's what we're focused on at Splice Machine, providing a relational database to run both existing apps, new Greenfield apps that people are imagining, Internet of Things apps, monitoring type apps, and doing that in a cost-effective manner. So when customers evaluate you versus traditional database products, what's the price differential? And then how does that figure in, you know, there must be other considerations in the choice? Some of it is, you know, and I don't want to discount the Hadoop movement. So some people will come to us and say, we are moving to Hadoop and we're moving our applications there, help us try to figure that out. But the ones that we're, I think, the most successful at are when people are making a decision between maybe Oracle Rack and transitioning to Oracle's Exadata product, which is a much more substantial product, but they're having that sort of challenge. And there's a cost-effect, you know, cost problem there. And we get to come in and use the open source community that we contribute to. And it gives us a nice firm foundation to build our relational database on. We use those as a nice cost savings for our prospects. And is the catalyst that they're making this move, so now it's a chance to kind of take a breath of air and see what the other options are? Or is it just because there's such a step up in cost that it's really a forcing function? I think both. I mean, I've heard multiple customers say, this business is not viable on my current database stack with the growth of data. So that's problematic. Not because the growth of data. The growth of data. And then they can't go back to their customers and pass on those costs, unfortunately. And then the people who are adopting these new technologies have a different price point. So it's a forcing function for the businesses. The cloud is a forcing function. People saying, I've heard Fortune 50 businesses now tell me, I'm moving everything to AWS in the next four years. And you think, how many databases, how many Oracle instances do they have? Probably hundreds to maybe even... Does that mean they'll replace the application database underneath it? Or does that mean they're building new apps? I think on a lot of this, they're thinking consolidation. They're wanting to do it in more of a manageable scale-out architecture. But I don't think their old apps are just going to go away. That's why we based our technology on SQL. SQL 99 is standard so that your existing apps can run. And we've certified certain vendors already that are application companies. Redpoint is one that we've publicly announced. That's exciting for us to get existing application companies. So if you're running these apps, you already know that we work on that and we're certified with those vendors. So that's one of our key strategies. What's this an interesting dynamic? On one hand, you've got this incredibly rapid growth, the rate of growth of data, right? This coming up, we heard of static EMC world data. Like 95% of the world's data was created in the last two years, whatever. It's growing quickly, which is different. And then you hear constantly of this kind of asymptotically approach running to zero with infrastructure at AWS with this constant price reduction. And now Google's coming in trying to kind of go round for round. So as you said, now your customers are kind of stuck in the middle seeing or hearing about crazy reductions in infrastructure costs but then at the same time seeing crazy increases in the actual demand and the actual consumption or generation of this data. And I think the expectations have changed. I used to think kind of cloud was a bit of a hype personally. And then now I've in the last six months, I don't know, something changed. And I see it in not just small or medium sized companies that are trying to be cutting edge and nimble. I see this in large companies saying that they're going to make this move because it makes them more nimble. It allows them a whole different host of possibilities. You said something interesting about customers on Oracle Rack, the cluster version of Oracle. Moving tech or evaluating exadata because exadata squeezes more out. Is that one of those points where they realize I can't afford really to go to the limit with Oracle and I need to consider? You know, I mean, Oracle's an amazing, you know, we have people on our advisory board. I mean, it's an amazing database product. The question is things are changing and the expectations of people are changing. They want commoditized infrastructure. The environment is no longer the relational database player. That's one key part of our story. We're the relational database, but our data is then accessible by any of these tools in the Hadoop platform. Machine learning tools, things of that nature, they can go at us natively and they don't have to move their data off of our relational database onto something else. We're all using the same storage platform and we're prioritizing our access to that storage. That's a powerful, when you start to get these products with all the investments starting to line up, you think, wow, you know, I can run my relational database. I can run my applications and then I can do some operational analytics on that and I can start to do this in a much more agile way. And businesses are under pressure because these, you know, younger businesses and businesses that are much more, I guess, better to be honest with information technology are putting that pressure on them and their expectations are going. And the data that's two days old is no longer acceptable. They want it, you know, fine, 10 minutes, okay. But, you know, they're starting to want to do more trickle feed type things that are much more real time. They want to do complex event processing in real time. These things are exciting for us at Splice because that really hits our key differentiator which is a transactional system on top of a highly scalable infrastructure. Yeah, and you touched on it briefly there but again another theme at a lot of shows really, is the workload, right? So it's horses for courses. What's the right workload to be on what infrastructure? Are there any particular types of workloads that you see are kind of driving this change? I think, especially this conference, it's really talking about people building applications. Not just, I want a parking ground for a bunch of data and maybe a data scientist kind of plays around with it. Not the data lake stuff. Yeah, yeah, they're talking much more on operational data store. We have real applications on it that you're wanting to provide real things for your business to move your business. I think that's what's exciting to me, especially this conference because you're actually seeing some of the leading companies in the world. Carter from Google had a great thing where he's talking about. Google, Apple, Bloomberg, how exciting. Top companies are using this technology and they're doing things that other companies are completely envious of doing. And our goal at Splice Machine is just to reduce that friction for them to be able to do this as well. You have something that may be a real lasting differentiator versus some of the other new SQL vendors. You're part of the Hadoop ecosystem. How does that change the, how do customers evaluate that against some of the other vendors? It's a good question. What we usually say is, do you want to scale up or out? Everybody to a team will say, we want to scale out. Then the question is for new SQL, do you want to scale up on their platform and their ecosystem? Or do you want to scale out on the Hadoop ecosystem? And all of them will say, no, we see Hadoop as a platform of the future. We want to put more and more things on that piece. Now, the thing that hits us a little bit, just to be completely honest, is sometimes the Hadoop platform is known for analytics. And that can be a bit of a mixed message. In H-Base, I think in general, it gets hit with that because people talk about analytics on H-Base. H-Base is really optimized for transactional type workloads or operational type workloads. So it's an interesting kind of thing we have to work through. You said you gave a session earlier today. What was your session? No, we're getting ready to do it right now. It's a SQL Smackdown, which is interesting because Splice Machine is a venture-backed company. And a lot of the people that will be on the panel are more open source and are doing it in a very different way. So some are non-transactional. So it'll be interesting. I think it's going to be interesting to see what workloads they think their database can do and where they're at. So I don't think it's supposed to be a computational sort of thing, but it's going to be interesting. It'll be a little spicy. We've got to get everybody excited. I'm doing that actually right after this. All right. Well, John, if you looked out like three to five years, where do you see the new SQL guys? Where do you see what today are the independent Hadoop distro vendors and what they're doing? And then the traditional database guys, I can't just keep sort of rowing up river. What does this Mellonish look like? Yeah, I mean, the traditional database guys are retired at this point, the first generation that gave us a huge leap. I just talked with Samir from Facebook who was a key component of Microsoft SQL Server, for example. They are kind of going around. Then you've got new SQL. I think new SQL will definitely go forward and start to gain more and more traction personally. The platform players, they're absolutely critical for our business. We run on all, I shouldn't say all of them because there's more than this, but MapR, Hortonworks, and Cloudera, they are absolutely critical for us. The infrastructure monitoring they supply, the capital and all the investment in Hadoop and HBase, it's amazing. Those companies, you know, they are very, what they've done is extremely impressive for us. Awesome. We all love theCUBE, which is good. We'll be at Hortonworks. We just had Mike on the other day, Rob Bearden on the other day, so that's great. So thanks for stopping by. A few minutes before you have your panel. And good luck with that, the shake down? The shoot down? The shake down. All right. So John Leitch joining us here with George Gilbert. We are at HBaseCon 2015. You're watching theCUBE. Thanks for watching.