 Okay, we're back live inside theCUBE, our flagship telecast where we go out to the events and talk to the smartest entrepreneurs, developers, executive geeks, entrepreneurs, whoever we can find to get the information and share that with you. We're here live in the Node Summit in San Francisco where Node.js is having its inaugural kind of geeky meets business conference. It's not peer developer conference, but it's pretty much the geeks building out a new industry. It's really cloud-focused, cloud developer, DevOps. And I'm John Furrier, the founder of SiliconANGLE.com, SiliconANGLE.tv. And we'll be launching DevOps Angle, a new vertical this week. So look for that on SiliconANGLE.com or DevOpsAngle.com. And we're going to cover all the angles of DevOps. And I'm here with Bruno Fernandez-Ruiz from Yahoo. Now, what's your title? I'm a vice president and architect fellow on the platform group. So vice president, so you're a top guru over there on the large-scale side. Todd P., Dr. Lucky Spin was the cloud guy earlier at Yahoo. He's gone doing his new thing. He just announced his startup. So Yahoo's huge. We're not going to talk about the Yahoo conversations that everyone else is talking about, which is new CEO from Stonehill, Massachusetts, my neck of the woods, or any of the other stuff. But we're going to get more into the geeky side of Yahoo. Huge infrastructure, a lot of volume, a lot of traffic, a lot of cloud, a lot of large-scale. Just to keep the lights operational, it's pretty complex. Can you just share with us some of the geeky complexity involved in what's going on over there right now? Yeah, so one of the things you need to look at is 21 billion pages daily. It's a lot of pages and there are only the likes of Facebook and Google and maybe Amazon that really drive that. And those are not just static impressions, but actually you have to compose a page. There's a lot of recommendations, targeting, personalization happens on the page. It's not just one little note on the stack. It's actually many layers on the stack that are doing analytics, that are doing ranking, that are doing search, that are doing content management. I'm putting all of that together for 700 million users. It's the complexity of running big, large distributed systems that can fail in any moment. Hardware failures are constant. Network failures are constant. And you have to decide for failure. So in this Node Summit, Node.js is the new black, if you will, on the developer community. It has a real time element to it. They talk about kind of being a programming environment for people who don't want to be a threading guru or low level programmer. What's your take on that? I mean, what do you think of the Node.js and the benefits of it and how relevant it is in the production environment? There's some use cases out there, LinkedIn's doing some stuff with their mobile app. How much of this is real? Where's the upside? How much work needs to get done with Node.js? Is it ready for prime time? Take your choice of that. Many questions there. Take your choice. Jump right in. So for one thing, event-driven programming we've been doing it for a long time. We've been doing it with Intomi Traffic Server that became Yahoo Traffic Server. We donated it to the open source community as a part of the traffic server. We sustain in between 75,000 and 250,000 connections with each of these nodes, which is super. So the event programming mode works. It's very complex when a programming developer from the actual programming standpoint not to completely mess up your head with it. For us, we had enough of this technology. As a matter of fact, all our recommendation systems or targeting systems, they follow a similar event loop programming whether it's in C++ or in Java. The issue for us for Node.js and Intomi.js was can we shift the computation from the client side to the server side? The reason this is important is because if you look at the last 40 years, CPU compute density has increased maybe a million times. Networks had only improved by a factor of 1,000. So the phone that we carry in our pockets could do what a computer could do maybe 20 years ago, but the network is a little bit better only. It's flaky, it reconnects, it keeps dropping packets, and the experience that you get is bad. So you cannot rely just on building apps. The apps are going to become disconnected at some point and the question is how do I build the apps so that they still work if they're disconnected? Our thinking is that we want the same code to run on the client on the server side. And that's what Node.js is interesting for us. It's not to solve all the infrastructure problems behind Yahoo that are very complex for search and recommendation. It's just for that stream layer that if it doesn't run on the iPad, it doesn't run on Android tablet, it runs on our servers and you still get an experience. Emerging markets in Indonesia, extremely important. What's the implications to you guys, have you? First of all, how much note are you doing at Yahoo? A lot, a little, dabbling in shipping? Quite a lot, I'm growing. We call the program cocktails. Our strategy is to keep growing, adoption, all the new properties that are being rolled out. They're going to come into the new stack. We're going to target what we call continuous experiences. It's not just PC, it's not tablet, it's not iPhone. It's something that you start one place, you move to another place, it's continuous. All of that is basically going to be based on cocktails. It's going to run on our Manhattan Cloud Platform. It's going to run using Mojitos, a JavaScript-modulated control framework, full bet on Node.js and JavaScript as a technology that will enable us to do this. Have you guys thought of taking some of this great code and expertise and bottling it up for as a product to enterprises? No, we've thought about making parts of it open source to create further community and leverage. This is a problem during the panel that we just had. There was a question about JavaScript frameworks. And it's a very good question that there isn't anything good right now that you can do client and server side. We think we have something good, we're going to open source it to try and grow a community around it. It's good for us and it's good for a lot of people that will be able to use it and contribute. In terms of enterprise software, we want to provide a service. We're going to provide a cloud-based platform as a service for these Mojits, those Mojitos, Lifestyle and Town Applications that people can create for publishers that likes a Forbes, Source, Interlink, ABC News and so forth. So really the consumer focus, not business to business. That's, they're publishers are business, but they're... It's kind of that whole consumerization of IT, right? It's happening. I mean, it actually is happening, although it's not just bullshit anymore, it's actually happening. That's the challenge that the big IT shops have, is that they really want a Yahoo-like environment without all the headaches that you guys went through. So that's what's interesting about Joanne, it's one of these new approaches, is that they want that web-like scale that you've had through brute force, built, and Facebook and Twitter. And so what I'm trying to tease out is what your key lessons were that you learned there from a development standpoint. Because a lot of these companies don't have that background. They don't have the background in developers. Most of them have outsourced it mostly and don't really... They have administrators basically that sit inside. So the challenge is the app market's exploding with developers like here at node.js what is available. So I was trying to tease out how real note is through this thing and I'm sensing that the developer uptake is good. How would you peg that uptake in terms of the developers? Would you say on a scale of one to 10, 10 being actually totally mature of one or two? Or three. In terms of the developer uptake, the peer uptake on developers, I think we're probably not at three. It's a very fast, very exponential growth. It's one of the hottest technologies that you can be currently investing on. It has a lot of promise. It has a lot of hard work to put on. We put a lot of hard work to basically harden it, to be able to run search, to be able to run lifestyle, to be able to run the products that we do. And we suspect that most enterprise houses, they don't wanna do that. The same way that they don't wanna run their own publishing systems. They don't wanna run their own CRMs. They don't wanna run an ERP system. They just go and buy it or somebody runs for them as Salesforce. We want to be exactly doing that. We build a good stack. We call it the Yahoo Publishing Platform. We want to be start offering these two other publishers and Node, JS, and Manhattan are that shim layer that will give them that connected device experience. So a couple questions then we can go on to other things, but Hadoop has been a big part of Yahoo's DNA with the whole background with Hadoop. Obviously, Hortonworks kind of a spin out there. Got a cloud error out there. But mainly Hortonworks, a lot of X Yahoo. How much Hadoop do you have in your infrastructure is the first question. And then the second question is to get some big time developers, what's cool at Yahoo right now? From a geek standpoint, what's going on there that you can say, hey, this is a cool project. This is some cool development to attract those kind of developers. So I do- So we still invest in heavily Hadoop. We still drive in the community. We drive in the commits. We work in very closely with Hortonworks and things like the next generation, resource manager, is fundamental to our business, our all our pipelines, data pipelines, our advertising, our recommendation systems, the front page, everything on Yahoo really runs on Hadoop. So that continues to be the case. In terms of the whole exciting technologies, right? So Node.js is definitely one of them. We work hard with it. We work hard with JavaScript. But a lot of other opportunities in other areas, Hadoop being one of them. The things like our infrastructure as a service, our virtualization layer. We have a private cloud that we're working on. Very exciting thing for people to work on. Search technology, even though we're not on the deep web crawl search, we still need to search and rank items, structure data, web of objects, and so forth. So people with an information retrieval background, extremely good place to be for very large scale problems. Things like edge technologies and network people, people who care about TCP and HTTP and networking and being as close as possible to the user. Apache Traffic Center was just mentioning we have the next generation version of that. Leading technologies, Sherpa, PNATS, which is our distributed non-SQL store. You know, we've been leading, we published the paper in 2006 and we have innovated, we have a technology that is probably among the best in the industry, together with the Cassandras and the MongoDBs that exist out there. And they can come and work in that technology. So in terms of technology, extremely exciting and large number of opportunities that developers can join. You guys doing any H-Base, a lot of H-Base? We use H-Base, yes, we do. Good experiences with it. I mean, it's very efficient, but hard to program. It's new. It's new, we've been careful about how people use it. We're wrapping it with an asset. You can only test the H-Base, none of IQ level. It's hard, I mean. That's basically it. It's pretty hard. That's basically it. We protect developers from being able to use it directly. We're wrapping the usage, but we have probably one of the largest H-Base deployments in the world. What do you use H-Base mainly for? Web crawl cache. Okay, cool. Okay, vision. Last question. As the social web becomes more and more mature, we're seeing things like, just this cloud space here, you know, the Heru-Google bought by Salesforce. Just a few years ago, they were a startup. A lot of the public cloud entrepreneurs in this area, large-scale PhD guys at Stanford getting graduating, you're seeing some maturity in this marketplace. And you guys have been a leader in that. What's your vision for this world of social web and the web? I really call it Web 2.0. This is really the Web 2.0. But we'll say Web 3.0 for the sake of another extra digit. But what's the vision that you see coming around the corner beyond today? So I will not comment on the product experience because I think that's become a Hollywood production game. Maybe you're good, maybe you're not. Technology-wise, trend-wise. Real-time web is changing. It's going to change the way we understand devices. Devices are always connected now. Better or worse, but they're always connected. The ability to shift the cloud to the device itself, the ability to become real-time connected, your device becomes part of the cloud. We'll start to see that the storage on the device is actually joins the storage on the cloud. The computer on the device joins the computer on the cloud. And what you end up having is IEO being transferred between these elements of the big cloud. It's data peers on a big network. You want to call it peer-to-peer, synced real-time web or something like that. So you're saying is systems programming is coming back into vogue? And it's hard. And that's why things I know, JS, are necessary for people to grasp the rocket. So creating a simpler way to not to screw up H-Base or real-time build. To expose the IEO access to H-Base, exactly that. Okay, all right, Bruno, thank you very much for coming on Inside the Cube. That was a really epic.