 Live from San Francisco, California, extracting the signal from the noise, it's theCUBE, covering DockerCon 2015, brought to you by SiliconANGLE Media, with special thanks to Docker. Now your hosts, John Furrier and Jeff Frick. Hi, Jeff Frick here with theCUBE. We are live from San Francisco, DockerCon 2015. It's the second San Francisco DockerCon, I think the third show overall, so we've been here for two days live coverage on theCUBE, really getting out, talking to the folks, giving you a vibe for what's going on, and like we always love to do, we like to get the practitioners, not the people building the technology all the time or selling the technology, but the ones actually implementing it to make a change in their business. So we're psyched in this next segment, we have Steve Hoffman and also Rick Fass, both senior principal engineers from Orbit. So gentlemen, welcome to theCUBE. Thank you. Yeah, so we were talking a little bit before we got started, blah, blah, two days we've been going hard. You had a session yesterday, give us a little background for the folks who didn't make the session, what were you guys talking about? So we were talking about our implementation of our microservices architecture. It was sort of one of the more successful DevOps attempts that we've tried in recent memory, and Docker was kind of at the crux of it. And hopefully it went better than some of the Docker demos on day one didn't, they had the demo gods got them. You've had some bad ones within Orbit's. So talk about what you're doing with Docker, how is it changing the way that you guys implement microservices? How is it kind of a game changer and why are you investing in it? It's really enabled handing over more control to the developer of their environment. Historically it's always been the developers who would write the code and operations would deploy the code, they'd also prep the environment. Docker kind of eliminates that barrier. They now provide the code running in an environment and all we have to do is sort of deal with things like capacity and infrastructure and things like that. And if they wanna go do something or try something or a new operating system and new version of Java, whatever, we just, they just do it in Docker and we don't really have to be too worried about it. It's also kind of like trivialized the whole idea of a release or deployment for us because there's only one way to do it regardless of what's inside the application. So it's kind of let us focus on more interesting architectural and kind of operational problems as well as letting us just like the developers just worry about the code. Right. And is it changing their actual behavior, freeing them of the burden of some of the stuff that they probably didn't really want to deal with or didn't really feel it was core to their productivity? It is. So Docker is just one piece and it is kind of the middle of everything but a lot of what we've done is taken 15 years, 15 years of process that's been built up and up and up that's made delivering for us more difficult. I mean, it's had, there's been good things that have come out of it but ultimately, we took Docker and we also kind of took the whole process and said, let's get the people out and streamline it so that developers don't have to worry about getting approvals to go to production and developers don't have to worry about all the paperwork and red tape type stuff. They literally just make the change, make the pull request and let it go. Yeah, interesting. We had Michael on from Lyft yesterday and he was talking about, they could have a new person show up at work on Monday morning, have their laptop to them by 9.30 or 10, a little introduction, take you out to the new employee lunch and they are working at one and they're expected to push something into production by the end of the first day. That was fascinating to me. Yeah, we're not quite there yet. This is still kind of a new experiment but so far it's looking very promising and a lot of the teams that we've shown what we've done are very interested in trying it themselves. Right, right. And Orbus is an interesting company because you guys were a really early adopter back in the day of web technology, self-service, people booking their own travel, very disruptive, coming out of American Airlines, right? If I recall. Delta United, Northwestern and Continental. Okay, the only one I didn't know. Everyone but American. Oh, America was in there. They were in there, okay, good. Yeah, I thought America was in there. But anyway, so it was super new, super innovative but now compared to like a Lyft or an Uber who really starting with Greenfield, you guys have, how long has it been? 15 years of legacy stuff built up. At the same time, the rate of change has only accelerated the crazy growth in mobile. So talk about kind of how you guys look at kind of new technologies and new application development opportunities. At the same time, you do have your own legacy stuff in place. How are you kind of managing this? Orbits for a long time has not really been afraid of new technologies. Some people would probably argue that sometimes we jump on them too quickly for even before we should. But if it fills a gap or solves two problems, solves problems, then it's certainly on the table to consider. You know, the, yeah, it's, sorry. You just take it all, right? Yeah, but even if you think it makes sense, there's so much stuff from a technology implementation point of view. We just think, oh my gosh, if you're sitting on that side of the table, you've got this rampant explosion in cloud as a new mechanism beyond test dev and into production, you've got shadow IT for people kind of doing their own thing, which just sounds like this kind of helps a little bit. You've got crazy mobility and the expected behavior of applications now on the mobile. And then, oh, by the way, this little thing, Big Data, which you guys were probably well ahead of the curve with the kind of information that you've been tracking for 15 years, both in user behavior and the application, as well as travel, et cetera, et cetera. So when you're just facing this, I always hope it's like driving at night in a snowstorm with your headlights on. You know, it's like everything's just coming at your windshield. How are you picking? How are you choosing? How are you prioritizing? So, especially with the rate of all this stuff's going, I mean, when we started doing this stuff, so much stuff has changed, and we haven't bought into one full ecosystem yet and all the stuff that we've been building. So we've been piecing things together, Docker, Mezos, Jenkins, Ansible, even Chef for provisioning. We've kind of pieced together something that works great for us, but is also every bit is replaceable. And the thing is, when you come to these conferences, there's so much stuff that they're talking about that's trying to bring it all together, but everything's changing so fast that there's no best practices yet. And so, the best thing we can do using all of these newer technologies is make sure we're using them in a way that once there's kind of a better practice, that we can adopt it quickly. Right, right. And also, all the buzz is usually kind of precedes, maybe production-ready or precedes, as you said, documentation and support and all these other things as the buzz is just huge, right? It just continues to accelerate. And then do you have like an advanced technology task force or something where you can go kind of put them in the corner and noodle with them a little bit before you let them out of the cage? There's really no pedestal of, you know, like architects that sort of command from on high. I mean, we have a, I mean, basically everybody that's worked there is a really good developer and we don't necessarily want to tell them what to do. Right, right. We just need to make sure that the things that we need, right, like for instance, every change in production needs a change ticket. It turns out that 15 years of process, everybody thought that change ticket meant an approval from a person. And it turns out that no only in certain circumstances do you actually need a person to approve it. In most cases, for instance, the project that Rick was experimenting with, it didn't require it. And so we were able to take something that was very manual and just turn it into something that was automatic and that was that. Right, the bottleneck wasn't real. The bottleneck wasn't real. We just had to ask the question. It was just historical. Right, right. Nobody thought to think of. Now is that because of the development processes change with kind of this whole DevOps cultures where the lines have really broken down between those that build and those that deploy? Absolutely. And we want that to break down, right? Because at the end of the day, the person that wrote the code is the best person to know when it's working and when it's not working and to troubleshoot it. And so part of our job is to get out of the way and try and provide them the tools and the infrastructure so that they can do things more quickly. Changes to try things out, troubleshoot. And then if they want to deploy in the middle of the day and they're fairly confident with it, we let them do it. And if it breaks, they're standing right there to take care of it. Right, right. It turns out when you give the ownership to the developers, they usually do a better job anyway for the fact that they actually own it. Right, right. So they wrote it. Right, they know it best. Even though they don't get beepers anymore, right? To go off at four in the morning. So you've really seen the benefits of the DevOps culture. You've seen really the benefits of pushing that operational piece back to the guys that actually write and deploy the code. I would say the number of developers versus operations people, it's probably a factor, I don't know, 10 or 20 to one. And so we don't want it, the operations people do not want to be that bottleneck, right? We just want to get them what they need as quick as we can so that they can do what they need to do. Right, right. So talk a little bit about DockerCon. Again, you guys are here. You presented, you know, how long have you been involved with Docker? Are you guys contributing at all to the project or are you just purely consumers and excited that it's going to potentially change your world back to the office? At this point, we're more consumers than contributors. We're probably pretty close to maybe jumping out of the contributor bandwagon. We've contributed on a few pieces that not directly Docker, but kind of in this whole ecosystem. So like our service discovery, we just console. So we've contributed heavily to console. We've contributed some stuff for deployments on Jenkins with Mezos. We've contributed a little bit there. Nothing directly to Docker. Okay. Yeah, I've got something that I've been working on that I'm looking open source that's kind of like one of these bridge things where you're not all in on Docker. You have to use kind of what you have now, but it lets you get going there. Okay. We'll see. Any ship date ready to go? Yeah, it's just a matter of putting in a form that's consumable. Yeah, well, you got to write up the documentation and examples and talk. We've been working with Docker probably since about this time last year, I think. Yeah, we were here last year at the conference and when they announced Docker Hub and Swarm was like not even a thing yet. And that was why we kind of went down the Mezos route because that was kind of the most advanced thing. And now it's nice to see that there's this open container alliance going on and more standardization. And that's just going to help everybody. And is that good for you? I mean, were you excited with that announcement? Is that kind of, you talked about putting in stuff that you can swap out later if you have to, but that said, does it feel good to have some level of standardization on some portion of the total that you can then build off or have a little, I guess, higher confidence level maybe that it's going to last longer. I feel less locked in at this point. I mean, I think we're all in on containers, but the fact that they did announce that yesterday, that definitely makes you feel better. Because like I said, I mean, I walked down this hallway, talked to these vendors and there's so much new stuff and they just keep changing so fast. And I'm like, oh God, we need to get rid of this thing we're doing and swap this, it changes so fast. And I'm already thinking of things that I want to get us and swap out for things that I've learned in the past two days. And Oritz has been a big user of open source technology like since the beginning, right? And you just, it's good to know that we're going to have all that interoperability and things are going to be pluggable and it's not just, you do it our way or you don't do it, right? Because that never sits well and you always want, the technologies that are going to win in open source are the ones that are just better, right? If everybody's playing by the same rules and company A does it better than company B, then company A's going to win and company B's got to step up their game and make it better and that's how good stuff gets made. Right, talk, dig in a little deeper about open source. You know, open source, as a consumer of open source and you do have this, things get forked funny and then there's the support and where do you play but also in terms of within your own developer community, you know, there's a lot of goodness that comes to people who are contributing to open source, right? There's a lot of, you know, kind of Juju in the marketplace are feeling good. There's a lot of peer, peer feel good. Where does open source play for, you know, is it just as a, to consume it or how does it really impact your kind of culture of development? So definitely as a consumer, we've made a lot of, Orbitz has made a lot of contributions to open source. So Graphite was originally developed at Orbitz which is still probably, I would say, one of the most heavily used monitoring visualization services out there, wouldn't you say? Yeah. You know, we're contributing for console right now. We have lots of other, you know, older libraries that are still in somewhat use. They're in use. Yeah, this would be like, for instance, which was a monitoring framework. It never really caught on as much as some of the other projects like Graphite. But, you know, the idea of contributing back, you know, it's not in Orbitz's business interest to like write a web server. We want to use an open source one, right? We want to focus on the things that make Orbitz Orbitz and selling travel and getting, you know, people matched up with, you know, so that we're sticking in front of them the things that they're actually looking for. So they get to go where they want. We get, we get the transaction and, right? And things like writing a web server or writing, you know, a key store or things like that. It's just, it's not a competitive advantage. And so using open source and contributing back to open source, it just, it helps everybody. Right, right. So let's talk about your business a little bit. Or like I say, Orbitz has been around for a long time, 15 years, right? Start out, I just need to book, I need to book a flight and I'm not going down to the lady at the corner anymore. But it's a very competitive space. You guys have been at it for a long time. Traveling, generals, a competitive industry. So what kind of cool stuff are you guys working on? How are you using all these technologies, big data to really, not so much build a competitive advantage, which is great, but to continue to enhance the experience of the person interacting with your application. So we're doing a lot of cool stuff. One of the projects that's kind of in flight right now is we have a culture of experimentation on our website and so at any given moment, there's hundreds of what we, A, B tests running on Orbitz at any given time and the goal is to, well, improve conversion but also improve the user experience and make the website easier to use. Right, right. So we run these experiments and typically in the past few years, we run them and then we let them run for a long time and then we go in and say, we run a bunch of BI, big data, long running queries to figure out how we did and so, but we want to scale up to the point where almost every feature, especially as close to the front end, starts off as an experiment and we want to be able to find out if they're tanking conversion or improving it faster so that we can actually have more experiments that's running at the same time. So a lot of the stuff we've been doing with the microservices in-house, we're actually using some of that platform in Amazon to actually process the user data as they traverse through the funnel on our website and actually process that data and be able to run analysis on it in real time and so if we deploy AB test on the website that's tanking in five minutes, we'll know about it and if it's vastly improving progression through the website or conversion faster, we'll know about that way faster than we did before. Right, and quicker. So do you guys use Hadoop? We do. We do. You do? Did you go to Spark Summit last week? We had a couple guys that did go. We did not go. We were there, it's just a crazy time. There's just so much development. Things are moving so fast. So we're getting the hook. I'll give you the last word. As the trucks pack up tonight, we pack up all the stuff and we're pulling away here from the Marriott. What's the bumper sticker for DockerCon 2015? Stump them. It's going to be a fun year. It's going to be fun to see what's coming out. I'm very excited about some of the new stuff in the security. One of the problems with Docker with the sort of immutable images has always been how do you pass secrets in, like database keys and things like that. And there's really not a good solution for that right now. People just kind of jerry-rig something together. And it's nice to see that Docker and other people, these other companies are all working together to come up with really good solutions and not just hacks. A lot of our pain points have been addressed. By the things that they've been announcing this time. And not just Docker or the other vendors and stuff too. Right, right. And it's either like ready to go or almost ready to go. And so it's going to be a fun year. It's a good one. And that was pretty quick. It's only a couple of years, right? That this thing's been going on. So the fact that they're rapidly addressing your concerns is either they're really lucky, they're really listening, or I guess you fit in the middle of the belt curve somewhere. Well, they're on to something. Yeah, I mean the conference went from what? 500 last year to 2000. Yeah. You know, they're scratching an itch. That means scratching. Great. Well Steve, Rick from Orbitz, thanks for stopping by, sharing some of the insight. Like I said, we love to get the practitioner point of view. I'm Jeff Frick, you're watching theCUBE. We're at DockerCon 2015 in downtown San Francisco. We'll be back with our next guest after this short break. Thanks for watching.