 Live from San Francisco, it's theCUBE. Covering Red Hat Summit 2016. Brought to you by Red Hat. Now here are your hosts, Stu Miniman and Brian Gracely. Welcome back, we're happy to have on the program a repeat customer. We have Patty Power Betfair who had joined us back at the OpenStack Summit at Austin. Richard Haig who's the head of delivery enablement and he's brought along this time. Steve Armstrong who's principal automation engineer. Gentlemen, thanks so much for joining us. Thanks for having us again. All right, so Richard, I'll start. You guys were the award winner for modernization, I believe it is. Tell us a little bit about what that means, the project that you guys worked on. So it's always great to win an award for the project because it just gives back something to the teams. They can see some outcome for the effort they put in. The innovation award for modernization is quite special as well. To win something for innovation is always great as well. It kind of not gives some validity but it gives a real nice feeling for the work that's gone into some of the tech that maybe other people haven't looked at or haven't done before and we've bought it through, we've landed it and to get recognized for that, absolutely fantastic. Okay, that's good. We need an award to go back. There's been some interesting times in Britain between Brexit and Iceland and everything else like that. So we promise we wouldn't dwell on that stuff too much. I had a feeling that subject would come up, thanks. All right, so Steve, you're on the more technical side. Can you bring this unpacked for us a little bit? Some of the decision criteria who were going through and what led you to the architecture that you built? So we were really just looking for something that had open source at the heart of it. We didn't really want to get stuck with proprietary software and really what we wanted to do was go for something that had open API. So open stack really was key to the decision there. What we didn't want to do was write our orchestration or technical orchestration to deploy our applications completely around basically something that could go out of date if a vendor changed an API. So that was quite key for us. Also with software defined networking, we wanted something that would scale as well. So we looked at bringing in new edge networks for that to help. So marrying that with open stack was really, really key for us. Yeah, we had a chance and we were at the open stack some of the few weeks ago in Austin. We're hearing more and more customers, larger deployments. What were the big milestones for you guys to not only decide on the architecture but that you sort of knew it was moving along, it was becoming in production, it was solving the needs that you had? I think we basically looked at some of the other customers that were implementing open stack at the time and we were looking for other reference cases as well. So I think we found some other customers that had the similar stack to us and were looking to do a similar thing. And really we just wanted something that was key and open and basically had open source at the heart of it. So I think open stack was really giving that kind of AWS like ability for the private cloud. Right. So Richard, I'm wondering you've got to go to the open stack show down in Austin, you're here at Red Hat Summit. Can you kind of give us a little comparing contrast what you saw of the show themselves, the community and what your weeks have been like? Yeah, so for sure this show is a bit more focused on of course the Red Hat Suite and some of their customers. Summit in Austin was, I guess, a slightly larger community-based event. I think they've both been very good events. I always like coming to these because you get to meet other users of that technology. And in fact, there's quite often the chats that you have with those other users around some of their problems and how they've sold them. They end up being the most of value. So both very good shows, both in the US it's been very nice to come back again. I would definitely recommend either of them for other customers of Red Hat or for people looking at some of those open stack community projects. Yeah, you were telling us off camera when we were talking earlier, you said that you've had a number of sort of community-based meetings while you've been here, meeting with other customers, other people using RDO. Talk about what are those conversations, what's going on in those conversations? What are you learning from each other? How do they structure them so that you're all getting value out of those meetings? So a lot of them are great because if you rewind a year ago when we were looking for what would be this next generation of infrastructure, there was a lot of concern around is open stack ready? Is it ready for the enterprise? Is it ready to be used in anger? What would you put the crown jewels on it and actually run production? And actually to come back a year later and start talking to some of these customers who a year ago were at the same point of the journey that we were and all of us in that space of that year have come back and said, yeah, it is. We're running production workload on this. We've got our tool chains aligned up to it. We've done upgrades of the software that we're using and it really does, it looks and it feels and it runs as a production ready piece of software. So to come back and have those conversations has been absolutely fantastic. And okay, to also look back at some of the little problems we've had to share some of the issues is helpful both ways because we may come across something we haven't seen yet, at least we now want to look out for. So it's a really valuable use of our time for that. Steve, any classes or demos that you've been digging into while you're at the show, what kind of key takeaways have you had? Well, I went to one of the triple low ones, which is the uptreem project for Retact Director, which is the installer. Today that was very good because it was looking at lessons learned and performance tuning that you could actually do to Director. We'd hit some of those issues when we were actually working on scaling out the infrastructure only a couple of weeks ago. So if only that session had been a bit earlier, we could have probably avoided some of those things. But to be honest, it's just coming to these events and learning and sharing information with other open source users that's really, really great. Yeah, we talked a little bit the last time. You guys have a fairly dynamic environment. I mean, you're dealing with gaming. You have different sort of trends and emotions going on in the marketplace or people. Talk about what does the old environment look like in terms of scaling and operations and what does your new environment look like that allows the business to do new things? I think the previous environment was very static. So servers were long-lived. Basically, virtual machines would be deployed and just left. And then they would be catered for. Basically, you would do it in place patching. With this new environment, we've moved towards an immutable model where you basically destroy the previous release, bring up the new release, roll it onto the load balancer. That's really the model that we went to. So it's more dynamic and you can essentially scale up and scale down as the teams require. So that's the real difference. There's the automation around it and the speed of change that we can get through to production. And how's the team adjusting to that change of technology and change of operations? I think one of the main things for them is just getting used to the point where they don't actually have to wait on a ticket. So they're very happy that they can get it at such pace and then it doesn't become a bottle and they could have blocked it further. Richard, any of the announcements this week that kind of meet what you're looking for, especially around OpenStack hasn't been a big piece of what's going on there, but anything they're doing now or things are giving you visibility towards the future that you took away from the event? Yes, there's been a lot of announcements, a lot around containerization, which we're less into at the moment. Specifically around the OpenStack side of it, the announcement that they're going to much more closely in line with the upstream releases is great news for us. So it means we can get the next features out much more quickly and consume that more quickly. So that was something certainly that was great to hear. Yeah, either of you, how do you manage that upgrade cycle? I think that we probably talked a little bit about this in Austin, but you look at how many releases on what people are running today. How do you manage that? How fast do you want to move it? How do you move it to the newer versions? I think what we'll be looking to do is really do an OpenStack upgrade once per year rather than once every six months, because if we took it every six months, we would basically be doing upgrades constantly. And so one of the things that we're actually looking at utilizing is not doing an in-place upgrade, it's doing an immutable upgrade. So we basically spin up a new version of OpenStack in parallel to the current version. And then because we have the deployment pipelines, we can change the API endpoint of the users, and then they just deploy mostly to the new OpenStack version. So that's something that the software-defined networkings allowed us to do, that probably native OpenStack wouldn't allow you to do if you're tied into the Neutron component. So that's been good for us. We've heard a lot this week about Participate. You talked about looking for open source offerings that you don't get locked into. How much do you guys get a chance to contribute? Is it something that you want to do? Is it something you're allowed to do in terms of code back into the community or just being actively involved in the communities? There's something we always try to do. And basically, we don't want to just be people that take from the community, we want to give it back. So we're actually looking at now that the dust has settled on the rapid implementation that we've done. We're looking at putting back everything that we've written back to the community. So we've got Ansible modules, for instance, that we're looking to commit back. And there's other things that we've actually done with Jenkins job builder, for instance, that we're putting back already. So we've already made contributions back. Yeah. Ansible's an area, we had one chance to talk to a customer on the retail side. Talk a little bit about Ansible, how do you use it? What do you like about it? What drew your attention to it? So I think one of the great things about Ansible is because it opens naturally in Python and Ansible's written in Python, you can easily kind of put a wrapper around the repeatable tasks that you do every day. So generally, if you're doing something day in, day out as part of the continuous delivery pipeline, you can just write a module that another team can then reuse and do the same way. So it's basically a good way of collaborating because we've got three teams, we've now went up to four across multiple locations. So they can basically reuse those modules and do it the same way so that they're not running snowflake processes. Right, right. I want to give you both the last word on here. I guess a couple of things is, number one, are there any technology pieces that you're looking for in the future? Number two, is the Olympics going to cause a big spike in your activity or anything else from kind of the gaming gambling side? We're more than the open championship. Oh, the typical things. Will we rejoin your, leave Europe again? Yeah, exactly. I think the Olympics show will be a great opportunity for us again and drive some customers to the site. In terms of announcements coming up, I think working with the Ironic project later in this year I think is going to certainly be of interest and certainly amongst the vendors we're working with at something we're spending a fair amount of time talking around. Yeah, we just really want to give our delivery teams a choice because they're deploying virtual machines at the moment. We want to be able to cater for bare metal or later on containers as well so that they can choose because there's no one size fits all. Sure. Well, it always feels like a sure bet when we've got Paddy Power Bet Share on the program. Thanks so much for joining us. We'll be back with more coverage here from Red Hat Summit 2016. You're watching theCUBE.