 Live from Boston, Massachusetts, it's theCUBE, covering OpenStack Summit 2017, brought to you by the OpenStack Foundation, Red Hat, and additional ecosystem support. Welcome back, I'm Stu Miniman, joined by my co-host, John Troyer, we're really digging in to some of the practitioners here on day three of our coverage. Happy to welcome back to the program a CUBE alum, not only that, a super user, and not only that, a co-winner, Patty Power Betfair, Steve Armstrong, Principal Automation Engineer, thanks so much for joining us, and congratulations to you and the whole team. Thank you, thanks very much. All right, so we've had you on the program, bring this up to speed, you know, where your OpenStack deployment going, where are you spending your time, you know, the event itself? So we've just recently, last year, merged companies, so what we're doing with our OpenStack implementation at the moment is, we're migrating all of our applications onto it from the merged company, so we're in the migration phase of the project at the moment, so we just recently, just after Christmas, hit the 100 applications onto the platform, Milestone, we're now up to around 200 applications, so what we're doing with it is we've got a single customer platform, which is the merged code base of the two companies, and then we're going to run different branding from it. So in terms of OpenStack, what we're doing is, we're looking to do an upgrade in the next month as well, we had the session earlier on today where we went through that, so hopefully that was insightful for the people that were here, so, yeah. Fascinating, I tell you, one of the, you know, there are many challenges with mergers and acquisitions, IT can be atrocious, I've worked with plenty of companies, if they're small, the parent company comes in, rips out the entire thing, and puts in a new thing, how's OpenStack, is that an enabler, does it, do you see it as a marked improvement, and any findings that you've got so far? Well, I think with OpenStack, it's very flexible because we're using it as the middleware for the whole platform, and so if we've got different storage vendors, we can just substitute it in, and then go to the OpenStack APIs to programmatically control everything, so it's really useful for us, so if we ever want to essentially use a new storage vendor, then we don't have to rewrite all of the self-service orchestration that our developers are using and interrupt them, so that's really, that's key for us in our business. It's interesting you use the word middleware, I haven't heard that word used in terms of OpenStack, but you mean the layer, literally the layer between storage networking, the raw infrastructure, and then the app on top? Yeah, so what we're really doing, we've created a self-service template that our development teams use, we didn't want multiple different ways for teams to create virtual machines, and basically go to the APIs directly, so what we've done is we've created a layer using ThoughtWorks School, where development teams fill in self-service YAML files with all the details that they need, and then they consume the infrastructure that way, so we're simplifying it and making it user-friendly for them so that when they're onboarding an application, they don't actually need to come to the infrastructure team, they can basically self-serve against OpenStack, so I think that's giving them that AWS or Google Cloud or Azure-like ability within the private cloud, and we've had to really change the way our business is set up to actually operate that, so generally what we've done is we set up different teams where there are more T-shaped teams, so in a T-shaped team, you'll have a network engineer, you'll have a storage guy, you'll have some automation engineers, someone maybe from a development background, and what we really did with it when we were building the pilot process, we tried to encapsulate all those different skills within the one team and set them up as a core team that would then go and build infrastructure and using best practices from each discipline. So T-shaped in the sense that the team is still cross-functional, what's the T of the T-shaped? So the depth of the T is really the deep-dive expertise, so you might have a network engineer who has a deep-dive knowledge in that, but what we're trying to do is expand the team's breadth, so the breadth of the T is really the other disciplines that they're learning as part of that team. Well, and congrats on the award again. Thank you. As they talked about the award, some of the description of why you got the award, they did mention the words DevOps and CICD, so you talked a little bit about org structure and changing your org and processes to do that, is that, now do you call that T-shaped team, is that a DevOps team for you, or how do you all look at it? We don't really like to use DevOps team because it's kind of a leading question, so it was really, what we tried to do is have cross-functional teams, so really DevOps for us, what it means is more collaboration between those teams. We've still got teams at the moment within our business that are looking after the heritage legacy stacks at the moment, but what we'll need to do at going forward in our business is bring those teams into the fold, because we've really had, I mean, essentially what we're doing at the moment, it's like Gartner by Modo, where you essentially have mode two where we're built this, we then need to take that to the next level and basically bring the people that have been looking after the other parts of the business because you need to maintain them while we're doing this new private cloud implementation, along on that journey, so we're running training sessions now for our network engineers, teaching them Ansible, scaling them up, so it's really exciting times, just bringing them on that journey. I actually think that's fascinating, because there've been a lot of talk about Gartner by Modo, type one versus type two, and how that isn't, and the word from the community and from the end users, that's not sustainable, so what you're saying is indeed you can organize that way, but you've got to bring the old teams in. Yeah, I think you can put names on anything, but generally that's what you'll do, you'll stand up, we stood up a brand new green field implementation, you needed people to go over to that and act in a different way, because OpenStack, it doesn't make sense having different silos looking after different components of it, because OpenStack centralizes that into middleware, so it's actually quite difficult to chop that up into different silos, if you're going to do it, you couldn't have someone just looking after Cinder for instance, because it's so incorporated with the rest of the stack, so really what we're doing is, we're exposing that API layer to the developers and allowing them to self-serve against it, and then we look after as the core team, the maintenance of it, so we've done this with a team of around eight people looking after the core platform, and then we've got multiple different teams that went out and they help the developers on board and migrate applications onto the platform by teaching them the self-service workflows and how to fill out all the YAML files, and then if there's any feedback from them, we use a continuous improvement model to try and get them to improve the platform continuously, so it's a continual ongoing process, and it gets better and better each time, and hopefully we're going to speed up the amount of deployments that we can do and speed up time to market for it. Right, so Steve, we've very much appreciated your organization sharing with our community, you're very active obviously in the SuperUser. Talk about how you interact with your peers, how that's helped with your learnings, kind of that give and take that you have with the community. Yeah, so with the community really, we come to these events and we generally try to be as open as possible and just talk about our lessons learned. I think OpenStack Summit's great for that because people are very honest, it's not like vendor-led meetups, for instance, where they'll just tell you that everything's great, and they're very self-deprecating in some of the sessions, but I think that honesty with the open source community and the continual learning that you get from that is really key to actually looking at the problem, saying, okay, we're not 100% perfect because you never will be and continuously improving as a community, so I think having the belief and the drive with the open source community is very key in that because I think what you can do is, if something in OpenStack isn't working the way that you want it to, you can contribute back and you can actually help make a difference and make it better, so that's what we're trying to do. There's projects such as Viterage for root cause analysis where at the moment, they don't have a Sensor plugin, we use Sensus, so we'd be keen to contribute back and write a plugin for that project so that we can use it and then others basically benefit from that as well, so I think that's where OpenStack's very key, you heard Edward Snowden's keynote, some controversial things in there, but at the same time, the premise was really, if you're putting your data somewhere else like in public cloud, you don't actually know what's happening with it, so that was something that resonated quite well because you have to look at what workloads you want to run in public cloud and which ones you can run in private cloud, so I think it will really, we're just getting onto the next stages and evolution and that journey where we will be looking at what workloads we place where and I think that's where tools like Kubernetes are really thriving because they can place workloads wherever you want and that's why the popularity's so high. I'm wondering if you could speak a little bit to your company's corporate culture that allows this movement, I think information's open, eventually the house always wins on these bets. Right, with so much information available. Yeah, so I think for us, the way that we've been able to do this is we've had sponsorship from CTO level and director level down and it's very hard when you're doing a grassroots movement of just engineers trying to do this from the ground up, you really have to have a company that believes in this philosophy and wants to take it forward. For us, what we really wanted to do was just create a platform that allowed our developers to innovate on it and just basically make the best tools possible for our customers. So your longtime OpenStack user, this is, we're now here in Boston, summits every six months, anything in particular about the mood of people, the operators here, kind of how you would like to see both, we talked about Kubernetes and you talked about different modules that you might want to see some activity in or just how you're seeing the future path of OpenStack, how would you like the community and the project to grow? Well, I think there was a lot of presentations on standalone aspects of OpenStack. So you have Cinder standalone for block storage, you have Ironic standalone. We use some of those projects to actually build it out. So I think modularizing it and allowing it to be used, that you might not want to install all of OpenStack, but why can't you install Cinder, for instance, to control block storage? And so I think that's really the future of it. People could take all of it or they could take different components of it. And I think that's what we're seeing in the community. People want to be able to install Cinder to help manage it and maybe not install Neutron or Keystone alongside it. So I think that's really where OpenStack is going. It'll be a modular microservice framework that makes it up and you can install the bits that you want and the projects that you want. We've also seen a consolidation of projects. There was also talk around essentially making projects simpler and removing teachers. I think when we originally had OpenStack, we just tried to throw every feature possible in and then you seen a sprawl of projects and then that's not maintainable. I think what we're getting down to is just the key projects that will then use going forward. So I think you'll see the consolidation and then standalone instances that you can kind of plug in around the edges. So Steve, why don't you speak a little bit about your business? I have to think there's few companies, at least definitely fewer industries that deal with the rate of change and the uncertainty in the world more than really gambling and everything that happens there. Anything changing in kind of the relationship of IT to the business? How does OpenStack help you respond to a very dynamic environment? Yeah, so I think the key thing for us is if one of our competitors has a feature and we can't compete with that feature, we just will lose our customers to that competitor. So really being able to change and use OpenStack to change the platform and get new products out to market as quickly as possible is the key for us in that. And generally OpenStack, it's helping us. We want an active, active data center. We have a 24-7 business. We really need to have that uptime. If we're down at any sporting events, our customers will go somewhere else to place bets. So that's really key. For us, we've used OpenStack across two data centers and built that out. And what we're looking to do is scale that out horizontally. So for instance, when we've got new applications coming on board, we can just scale out new racks using OpenStack. We use Ironic. We're completely controlling the whole data center programmatically. And that allows us the ability to scale up the infrastructure to meet the demands so that people are not waiting on tickets or not having the internal IT processes that are hindering most other firms. So that's really where OpenStack is allowing us to evolve. It's that flexibility in having a private cloud just like you would a public cloud with AWS. But we've got that in-house. So I think we're quite lucky. And I keep telling the graduates that are working on this. This is a once-in-a-lifetime project. And I don't think they'll really believe me until they get their next job. So I think they've been quite spoiled in this as well. Steve Armstrong. Really appreciate you joining us again on the program. And once again, congratulations to Petty Power Betfair and the whole team. And John and I will be back with more coverage here from the OpenStack Summit 2017 in Boston, Massachusetts. You're watching theCUBE.