 Okay, welcome back to New York City. We're live here at siliconangle.com, siliconangle.tv, it's theCUBE. We'll go out to the events and get the signal from the noise and we are here with some, a lot of signal with two great guests, friends of ours and the big data space, pioneers, who have been in stealth mode for a while but now are out with a public launch of their company and product. Todd Papianu, the co-founder and CEO of Continuity and Jonathan Gray, co-founder and CTO. VP of Engineering, what's the title? CTO. CTO. Welcome back to theCUBE. Thank you very much. These guys have been on multiple times. Certainly Todd and Jonathan. Jonathan, you were at Facebook when you were first on and Todd, you were at EIR at Battery Ventures. Now you're both, you've been in stealth mode and all the time, it's kind of dancing around, talking about what's going on in the industry. Now you've launched your company. So one, I want to get this out there for you guys to talk about Continuity because you guys did launch your company. And what's notable is, Jonathan, you're at Facebook. You know about Big Data, you know about Hadoop. Todd, you're a chief architect at Yahoo and this is the web scale companies that invented the whole concept of where Hadoop is going today and everything happening in the industries revolving around the web scale companies. Yahoo, Facebook, who essentially set this ball in motion. So congratulations and thanks for coming on theCUBE and being specific. So let's get it out there. You launched Continuity, Todd, you're the CEO. Let's just get this out there. Tell us about the launch, the company, the positioning and the product and we're going to do a product demo. Yeah, yeah, we launched a company just a couple of days ago. It's been a pretty steady storm of media press, analyst briefings, things have been going really well. Got a booth here at Strata and we had a ton of people come by the booth. At the company Continuity, we're a developer focused application fabric. You can think of us as the next generation application server for Big Data applications and we have this Continuity app fabric layers over the top of the underlying Hadoop infrastructure and we really give you an application runtime and container to actually deploy all the applications into. Our goal is to really enable the next wave of application development in the Big Data space. We think that right now, and I just saw Abhi talking about it, there's a lot of low-level infrastructure, but there's really not many apps. Abhi's got a nice app, but we need a wave of explosion of apps in the Big Data market. Without apps, no one's going to really do anything. So I want to just pivot off that real quick because Michael's last year with Ping Lee said, hey, we're a $100 million fund. We want to promote a lot of apps and one of the things we've been talking on theCUBE yesterday and this morning was, that just kind of didn't happen. Now analytics is a killer app, no doubt about that. That is happening, insights are great and that's going to continue to grow, but really on the application side, you really haven't seen anything. It's just too early. It's too hard, that's the problem, right? It's too hard, too low-level for people to build apps. Tell us why that didn't happen and we'll go right into more detail on that. Yeah, I think there's a huge barrier, barrier to entry right now with building apps, right? All the infrastructure's super low-level, getting out, infrastructure up is hard. You can go to Cloudera, those guys do a great job in packaging it, but even once you get it up, still the APIs are kind of like data kernel-level APIs, right? Most application developers want to work at a high-level set of abstractions, right? That's really a barrier to entry to actually building apps and even if you get an app built and do your two to three month kind of like, you know, science project, then once you run the app, monitor it, manage it, there's no tools out there for doing that at all, so that's specifically- No tools at all. Well, that's what we say anyway, of course, right? There's one new tool now. It's called the continuity app fabric. We're going to get to that in a second. That's an exciting product for developers. I mean, MongoDB's been very successful because it's been easy to use for developers. In the developer community, going back to open source, LAMP stack, I mean, spin up instances, what a concept, really easy to get stuff done, at least test and launch apps from a local server too, so that's essentially what you're going for here, right? That ease of use concept, is that true? Yeah, I mean, I think if you look back into building Java server applications 20 years ago, 15 years ago, it was really the application server and the creation of a bunch of higher-level APIs, reusable building blocks, you know, allowing individual developers to stand on the shoulders of everyone who came before them, who provided frameworks that made it very easy for them to focus on their application logic, their business, and forget about all the minutiae of the infrastructure and low-level stuff, and that really hasn't been done yet in big data. Yeah, I mean, I think one of the things we talk about in theCUBE, and we're passionate about, I'm a startup junkie, we love startups, startups create the future, and developers are the key, key equation to that in terms of value, even in the marketplace, so if you don't hit the developers aren't rocking them with product, there's no end game. So this is a real, has been stunted the market a little bit, I mean, I know this has kind of saved the day, but explain some of the things that you guys are doing, helping developers and the vision and how that renders itself with the product. Yeah, I mean, I think there's, you know, we focus on the horizontal application development lifecycle, right? We want to take a developer from my dear creation all the way through writing, testing, debugging, figuring out locally, then pushing onto, you know, up into production. We give people tools to, you know, plugins and SDKs and code samples and everything, you know, with a local version of our app fabrics that they can build our applications locally. When they're ready, they push a button, they ship it straight up to our cloud and they just run the end production up there and then we give them tools for doing the DevOps context up there as well. So before we get into the demo, Todd, I want to ask you another direct question. Obviously, we're big fans of HBase, we've been playing with it. Explain to the folks what's under the hood because it's really not that, I mean, it's so powerful, okay? And, you know, we've seen direct examples of the power. You're going to show a demo which I've seen, it's actually really easy to understand and use, but I want you to explain what's under the hood and why HBase is so powerful and why it's so hard at the same time. Yeah, I mean, so our entire platform is built on Hadoop, HBase, Yarn, ZooKeeper, the whole menagerie of the Hadoop ecosystem. And that's part of the challenge I think of people building applications today, which is you're not just using Hadoop. If I want to do random access, I need my HBase and if I want to serve, I need a caching tier and if I want to do streaming analytics, I need another system. And so people are building this patchwork. They're gluing it together. They're gluing together a whole bunch of different infrastructure components and what we're trying to do is say, okay, well, let's put all this stuff together into a unified platform and let's layer a scale out application server layer on top of it. And then selectively expose the different functionalities of all these different infrastructure components into our APIs. And so we attract away from the very low level Hadoop input formats and output formats. HBase, as you know, has byte arrays and very low level APIs around reads and writes. So designing a schema on HBase requires really six months of understanding what HBase is. And so what we do is instead of- They might have to schema change or are they data source changes? Yeah, yeah. And so, you know, instead of- New library, another six months. I think instead of giving people low level infrastructure which assumes no use case, we assume some use cases and instead give people really easy ways to build search indices or build counters and OLAP cubes or if they want to build time series databases. So right now you can do that with HBase, you can do that with Hadoop, but you have to actually understand how to do that, right? And so by giving people object models higher level things, we give people access to a bunch of really cool patterns that can be done on top of HBase but we implement those patterns in a really efficient way and instead they just get to use the objects. Yeah, one of the things, I mean, you can extract the way the complexity that's a win but it's also hard as you mentioned but also developers need an ease of use environment. So simplicity and flexibility is key. What did you guys do there around the simplicity and flexibility piece? No, I think one of the things we focused on, right, was like right now the developer experience in the big data industry is, it's a bit like the homebrew computing club experience so the computer's back in the day got a memory card from here and a disk drive and keyboard and you sold it all together. What developers want is really more of the Apple or Visual Studio experience that's nicely tightly integrated, works together. So we took all of this underlying infrastructure, led our app fabric on top, as John said, exposed the capabilities and did it through a unified API and we give tools to basically beautiful abstractions and beautiful interfaces as our kind of like key motto internally. That's what we're trying to do for developers. Awesome. Well, first of all, we're big friends of you guys and your work. I think it's a great product. We think it's going to hopefully explode and enable more developers to actually do more general purpose apps. So let's set up the demo. So which one of you guys wants to set the demo up as context before we start driving the demo? I'm going to run the demo here. Okay, so we're going to go to the camera and you're going to have voice over on the camera so let's go take a look at it. All right, cool. Are we ready to go? Yep, yep. All right. Let's hit it. So what I'm running here is the thing we call our single node addition, our developer addition of the platform. And so we've actually built an emulation layer on top of Hadoop and HBase and a single node version of the platform. So it's very easy for a developer to run this on their laptop and desk, debug, profile and do all that kind of stuff. What I'm going to show you here is creating an application, deploying an application and then running that application. So just going to create something called the demo app. Right now there's nothing in it but I already have my recipe prepared here. That would be in the local machine, right? So we just have that. Yeah, so all of this is running on local host right now and then at the end I'll show you how we push it up to the cloud. And so this is an application that's made of three different components, two flows and a thing we call the query provider and I'll show what each of these things look like. So just drag and drop to deploy that into the platform. And the key thing here is of course, it's just Java files. Java developers are building Java files all day, every day. They just package it up locally in their IDE. It's drag and drop ready to go. So we've got our Java files deployed, they've been exploded and deployed into the system. I'm going to start everything up here. What we see up on top here is two streams. Streams are how you get data into the system and so these turn into REST endpoints allows you to stream one event at a time or in batches all of your raw big data into the platform. Flows are our real time stream processing engine and so this green Laws and Dry here is a stream that's the input into this flow. And I'm going to start my driver script here, which is making REST calls into this stream to feed it data. So that's a simulation file, you just test it and look it in half. So just put a demo. Got it. This way we don't have to use the internet. Yeah, yeah. What's your favorite? Internet, the conference has been a bit dodgy. Yeah, yeah, of course. So what we see here is what we call tuples, individual events flowing through our flow. Each of these individual circles is what we call a flowlet. And if I pop this open, this is our flowlet pop up and it gives you more detailed insight into exactly what's happening inside of each of your individual components. So for example, how many data operations we're doing, how busy it is, the number of tuples we're processing. One really cool thing that we've done with Bigflow is it's elastically scalable at runtime. And on our UI, it's as simple as hitting a plus button. And hitting that plus button, clicking okay, there's now two of this flowlet running. And so on the local. So you're adding resources at that point. Yeah, yeah. So in the local version, it's spinning up another thread. Up in the cloud, we're using Yarn to actually deploy a new VM, deploy a new JVM, run a new thing. And then in the back, we're rewiring the queues and doing all that other stuff. But to the developer, they click the plus button, they hit okay, and everything just happens automatically. Application stays running, never goes down. That's part of the key for our platform, right? It's like we take care of a lot of the hard work. Like everybody in the continuity team has built some of the biggest, baddest, big data applications and infrastructures on the planet. Places like Facebook and Yahoo. We know what we're doing. We want to enable developers to be able to do it very simply and we take care of the hard work forum. I think the average developer doesn't want to get in the weeds and know what's going on with Yarn, and obviously it's evolving. So it's just a great service. I mean, it should be easy to use. Yeah, I mean the goal is to surface what developers care about and to try to abstract away everything they don't care about. So what's going on now? So this is a list of our data sets. And data sets are something we've built on top of HBase and Hadoop that is what I was describing before. We expose patterns instead of raw HBase. And so these are all a bunch of countertables and I think we call a sorted countertable. And so this is just metrics around how many writes am I doing, how many reads am I doing, how much storage do my tables doing. Last thing I wanted to show real quick here is our queries. Queries are analogous to stored procedures. This allows you to deploy another JAR file, which is one of the ones I deployed earlier, that contains a request response method. You, we bind it to a REST URL and shuffle all of the requests in there. And so in this application, this is actually one of our customer's applications. They hit this REST endpoint from their web tier to actually serve data. And so a lot of the applications we're targeting initially are much closer to the serving tier. I think that's one of the key things there, right? It's like, you know, you talked a lot about analytics being the key app. We actually think that there's a whole ton of kind of like new data applications, closed loop data applications that people want to build. So we're basically trying to take, you know, offline the batch out of it and you take signals in from, you know, from your app, you process it and you push signals back out, right? We're trying to build these real-time closed loop applications for, you know, consumer intelligence apps, as we call it. So a lot of people have been complaining about the cloud as they're going to rather skin a bare metal because of HBase and memory and I looked at Amazon and even Google Compute Engine, people who have been kicking the tires there, like it's just too costly to run stuff in the cloud. How do you guys talk about that and how does that relate to you guys actually pushing the cloud? Yeah, so, I mean, so, you know, John's been showing the local single node developer suite and as a developer, you're writing your code and you're testing it locally, debugging it in your IDE. When you're ready and you've proven that your app actually works, we have a very simple capability here where you just basically push your application up to the cloud. What we do is we take all of your application logic, all of your configuration and we deploy it actually up onto our private cloud. So we offer that in two flavors. One is we can host and operate the entire private cloud for you. It's a full HBase Hadoops, you know, infrastructure stack plus our stuff on top of it. Or you can take our private cloud edition and install it over your Hadoop and HBase installation if you have one already. So yeah, so I can build my own cloud. So to you, cloud means provision. It's really a pass, right? I mean, we think of, you know, private pass, private cloud, that sort of thing, right? And we're trying to raise the abstraction level, right? So it's not a requirement. Cloud is not a requirement. You're just looking at it as an endpoint. Yeah, I mean, I think we really think about it as a developer pass. Wherever it's deployed, it's a developer pass, right? To the developer, it's very much platform as a service. They're not worried about the infrastructure. They're not worrying about how many nodes and all that kind of stuff. To them, it's a service. Awesome. Guys, final question for you guys is, what's next now that you've got, you've been funded, so you got some great big VCs, Andries and Horowitz, Battery Ventures. Ignition. But I miss anyone, Ignition. Ignition, great, great firms to tier one VCs. You guys are obviously a rockstar team. What's next, you launch the company. Is it sales? Is it ramp up the engineering? All of the above? All of the above, onwards and upwards. You look at our website, World Domination. Right? Okay, so we have Todd Papayano, Jonathan Gray, both technical leaders in the industry, and Ta Lo, Todd is the CEO. Great to have you on the Cube, moves my mic here. We'll be back with our next guest after this short break. Continuity, check them out. Platform as a Service. Hopefully increase the developer traction and get better apps out there. We'll be right back after this break. Thanks a lot.