 Welcome back, we are here at Perkota Live 2014, Santa Clara Convention Center, Santa Clara, California, the heart of Silicon Valley. I'm Jeff Frick, you're in theCUBE. We take theCUBE out to all the great events. We get the smartest people in the room. We ask them the questions you'd like to ask them, extract the signal from the noise and really bring you the story if you weren't able to get here. So I'm joined here in theCUBE by Alex Yurchenko, co-founder of Codorship. Welcome to theCUBE. Thank you, Jeff. So you've got a really interesting product name. I want to jump into Galera Cluster. And I was struck by the name because a close friend of mine, Portuguese buddy from Brazil, used to say, we are Galeta. We are Galeta. We are a kinship, you know, a group of guys, a group of five or six guys. I'm just curious if that has any play into the name of your product. Well, it's definitely a Roman word and it must have some connection to folks, to people. But initially where we started off with was that Galera is a Roman word for what's in English is called a galley, a ship with multiple oresmen and maybe multiple even rows of ores, you know, Privyamas, Privyamas and stuff like that. And why we choose this name is that first of all, oresmen must work very synchronously in order to not interlock their ores. And secondly, and Galera is a synchronous replication solution. And secondly, the more oresmen you add to Galera, the faster it goes. And we were aiming for scale out and the more nodes you add to the cluster, the faster it goes. And ironically, or maybe not ironically, anyway, this synchronous operation is exactly the thing that allows us to scale up pretty much like in Galera ship. So synchronous and able to scale. Just add more rowers to the boat. Yes. As long as they're in synchronized fashion they'll make the boat go faster. Yes, exactly. All right, well that's a good picture. So let's talk about the technology. Where is this being used? What exactly have you built on the technology side that's different, that wasn't there before? And give us a little bit of background. Right, I've been waiting to say this. Here we are, we're breaking it live on the queue. Yeah. For many Galera users and just people who are familiar, have heard about our technology, it's just a replication solution for Maya scale. Not that sexy. Well, it's not that sexy, but in reality, Galera is a library, is a generic library which implements an efficient way to serialize access to a global resource in a high latency environment. Okay. To put it in more concrete terms, global resource may be a row in a table in your database. High latency environment is pretty much anything that is not in FiniBand. Anything where global locks are too expensive to obtain. And that's pretty much the whole world. If you want to have concurrent global access, write access to your data, Galera is pretty much the most efficient way to achieve that I know. So it's about knocking down latency in what you call, what you define as the high latency world. Yes. So that's interesting, because all we ever hear about is less latency, less latency, really driven by the mobile first. You get the Google developers at any of the shows that they're at and they're talking about, just shaving, shaving, shaving, bits of time here, there and everywhere. So what are some of the, I guess, legacy environments that are high latency where you've got an opportunity to go in there and trim? The applications and the customer use cases. Well, just this Monday, J-Pipes gave a very compelling use case for Galera in OpenStack, where you want Keystone, Keystone is a part of OpenStack infrastructure, where you want Keystone data to be known globally. In all your data centers, you want to distribute and you want to have full access to this data. So Galera is the perfect tool for this job. And basically I think it's the only solution that I know that allows you to distribute Keystone data across your data centers. And the Keystone data in the OpenStack is really the key piece again, taking the great naming convention. It is the stone that holds the arcs together. It's what makes it all hold together. Awesome. So then you play in the OpenStack then obviously as well. Well, many OpenStack installations have adopted Galera. Okay, great. So talk a little bit about the company. Where are you in terms of the company, how big are you guys? Where are you based? And talk a little bit about having kind of an open source component in your company. But then how do you make money? So obviously you got to have some non-open source stuff or services. Give us a little background on the company. Well, perhaps I should start a little bit with some history. Sure. That our company is called Codorship. And this is because the founders, me and Sepe Yacola and Temo Olaka, we met 10 years ago in the company that was then called Emic Networks. And now it's called Continence. All right. And due to certain events, we left that company and then we decided that we wanted to start our own company. Of course. That's what we do here in the Valley, right? And we kind of, we're thinking that we, you see, we started this work like 10 years ago. Everybody has experience in implementing database replication solutions in Emic Networks. We started with the M cluster. Then it was Uni cluster. Then we worked with Continence on Tungsten and kind of, we felt that we have accumulated enough experience that we know how this should be done. And we wanted to implement it. So what was the whole, what was the whole that existed that gave you guys the thought that you could go out and do it a better way? What couldn't be done that you can do now? I think... Or was it just a marginal improvement in the process? No, it wasn't a marginal improvement. It's, our idea was to, you see, we saw how complex replication solutions are. We saw the deficiencies and we were thinking that we now know how to, you know, pretty much make a qualitatively new solution which replication that just works. Unfortunately with replication, you know, replication is never simple. But I believe that we have achieved our goal, say, we have only two mandatory configuration variables in our cluster. Two? Yes, so to set up the cluster, you need to configure only two variables in MySQL configuration file. Wow, so talk a little bit about hyperscale and, you know, big scale, big data. Obviously, you know, there's a lot in Hadoop and there's a lot in the NoSQL databases but a big topic of the conversation here at the show is, you know, making MySQL better for large volumes of data. Here you're doing replication that would only seem to make your job more complicated, more difficult as the front end of that pipe is getting bigger, bigger, and faster, and faster. What does it mean for you guys? What's the opportunity? How's it impacting your customers? Well, to be honest, we haven't yet tapped into that Hadoop pool, but we're starting to get some ideas how we can improve, say, Hadoop, for example. Hadoop has a name node and currently there is no kind of good solution for high availability for that node. It's pretty much single point of failure and there are some workarounds for that but, say, what we can do with Galera is make a multi-master cluster out of Hadoop's name node. Well, these are kind of plans or some ideas that we're having but it's all... Still pretty new. Yeah, depending on how much resources we can gain, it all requires some work. So, even though Galera kind of is not exactly targeted for big data, even there we can... There is an application for that. Sure, because it's coming, right? Big data is coming, it's everywhere. Big data is coming and what's, like, one year seemed impossible. Right now we have already clear ideas how to do it, sharding and stuff. Okay, so why are you here? What's important about Percona Live that brought you here to the show? Users and partners, we're here to talk to people. Okay, and how's it been going? What's the vibe on the floor, Howard? What are the feedback you get from customers on kind of the state of the union and some of the partner community? It's been very wonderful for us but, you know, it's kind of... It's been very wonderful for us for probably the last half year or so. People are getting aware of the technology. They are trying it, they are building expertise. They are using it in solutions, migrating from proprietary solutions to Galera. So it's not just this conference, it's the overall process what is happening. In our business. Sure, and will you be at OpenStack Summit? Yes, we will be. Maybe not me personally, maybe, but we will have a boost there as well. Because we'll be there, the queue will be there in Atlanta this year, closer to Finland than in San Francisco, maybe, because it was important last year. So that's good. So what's the next big challenge to overcome? What's your next big kind of technical or business objective you're trying to achieve? Well, there are certain limitations that Galera has and that we resolve to solve by the end of the year. But these are kind of usability problems and what people have long asked of us, for example, support for arbitrary large transactions and stuff like that. But I don't see it as an next big thing, it's just, you know, finishing the product. I think the next big thing will be sharding. We are working on design of a sharded Galera cluster that's pretty much addressing this big data challenge and we are trying to address it in the most optimal way. And it's, this very thing is actually, the challenges, what kind of keeps us going. Yeah, it's very exciting, yes. I can't wait to start working on that. But you know, there are users, emails, like mailing lists, negotiations. I mean, when I have really very little time to actually code. That happens when you grow business, right? You have a business instead of writing the code. So is that, is the sharding project, is that a three month project from now? Six month project, year long project? It will start probably in a year. Start in a year, okay. Yes, I think, well, we will probably have some alpha. Okay, so we're here next year, Percona Live 2015, we get you on theCUBE. We very much hope to have something to show. Okay, good. Well, good luck. This sounds like an honorable objective and certainly a fun one to keep you working and keep the team going. So again, thanks Alex for coming on from co-founder of Codership, the Galera cluster. Best of luck to you and the team. Thanks for coming on theCUBE. We will be right back from Percona Live 2014. We won't be at 2015 for a year with our next guest after this short break. We're watching theCUBE from Percona Live.