 Live from Vancouver, Canada, it's theCUBE at OpenStack Summit Vancouver 2015. Brought to you by headline sponsors EMC and jointly by Red Hat and Cisco with additional sponsorship by Brocade and HP. And now your hosts, John Furrier and Stu Miniman. Okay, welcome back everyone. We are live in Vancouver, British Columbia for OpenStack Summit. This is SiliconANGLE, Media is theCUBE, our flagship program. We go out to the events, extract the signal and noise. I'm John Furrier, my co-student Miniman. Our next guest is a user, a customer of OpenStack, rolling it out. He's got scar tissue to prove it. Nick Gerasimados. That's correct. Director of Engineering and Cloud Search at FICO. Welcome to theCUBE. Thank you. So we love talking with users practitioners because, one, it's emerging, it's exciting. A lot of new stuff in there, but there's a lot of opportunities to create a lever of innovation. Agreed. Share your story. What are you guys doing? What was the big picture? OpenStack, I'll see private cloud initiatives. Share a little bit about what the environment is, what you guys are doing, what was the business objective, and we can jump into it. Okay, so traditionally, FICO has always been a very on-premise provider. So we would always deploy hardware at a lot of tier one banks, financial institutions, government institutions, so on and so forth. And what we realized is that we were actually missing out on a very large, small, medium-sized businesses, financial institutions, and so on and so forth. So it was imperative for us to kind of embrace a cloud strategy. Not every single, small, medium-sized business is going to actually want to invest in hardware in a data center and have the operations to actually support that design. So I think that's really what pushed us towards a web-centric model. So talk about the challenges. What was the main reason why to go with OpenStack? Talk about the, and the rationale behind it, and then what were the first couple steps you took? Okay, so our legacy environment was based on VMware, a very VMware, Cisco, UCS-centric environment, large, blocky, if you think of like VBlock and FlexPod and things along those lines. Heavy investment up front, and then it trails off towards the end. What we needed to do was be a little bit more agile as we were going global. So that kind of pushed us to go towards the whole OpenStack design. Low point of entry, scalability, software-defined networking, software-defined storage without having to pay the penalties. Yeah, Nick, so can you walk us through a little bit? How many people were involved in this project? Roughly size of budget, if you can share that, how long it took, and gosh, how many pieces along the stack did you need to change? So we adopted it aggressively. The RCIO, Tony McKibber, essentially came to us and said that we needed to develop an OpenStack solution and implement it within a 12 month timeframe from start of nothing, bare bones to actually a fully functional solution. And that's both the infrastructure, the service component, and the PAS component as well. So it was a lot of work. It was fun, but it was also stressful. Having to validate different applications, different products on that same platform, and some of these applications and platforms were never actually even tested on a traditional legacy virtual environment. Big sand, things along those lines. So all in infrastructure and platform as a service, talk to us a little bit about kind of before, blocky, VMware, afterwards, your Red Hat customer, use an OpenShift. So was it new people running it or retraining? What's that transition look like? So we brought in kind of a core group of, I would consider them to be a SWAT team or a technical team, where we executed a lot of the vision when it comes to that. We looked at a lot of different distributions of OpenStack, we settled on Red Hat because we have a very strong relationship with them. We were actually one of the main customers that deployed OpenShift, and so we wanted to continue that. I think it's been mutually beneficial, specifically because we're not just consumers, we partner with them heavily. So what was the, I got to ask you, what was the biggest thing that you've learned? We could share it to folks out there watching, because everyone is going down, a lot of new people coming into the journey of OpenStack, the sessions are packed, and there's a lot of great technical stuff, Red Hat, everyone's got this good stuff going on. What's the magnified learning? So you can walk away and give three pieces of advice. What to stay away from? What to double down on? What did you guys learn? So one thing we did learn very early on was that CEPH, while it's a great platform for software-defined storage, it doesn't necessarily fit all of our needs. Applications that are very latency-sensitive or IO-sensitive, lots of transactional stuff. We realized pretty early on that CEPH wasn't going to provide the performance that we needed for that. So we brought in SolidFire from that design perspective, to guarantee SLAs, guarantee performance. There's other reasons why we brought them in as well. The other thing we learned pretty quickly was that Neutron, at least in the older versions of OpenStack, was not quite ready for production. So that was something we had to learn through very quickly. And then the upgrade cycle. So I'm sorry, just to poke on that, so what did you do since Neutron wasn't quite ready for your production environment? Well, we looked at Open Daylight, we looked at other different SDNs. The end solution is we did go with Neutron for our design, but we also aggressively moved from OpenStack version five to six, and we're also preparing right now to go to version seven. Neutron's grown a lot in the last couple of years. It's definitely much more stable than it was in OSP4. Yeah, absolutely. We saw it taking some of the ODL pieces and moving them into Neutron. So knowing what you know now, what advice would you give to your peers as to when putting things together, did that SWAP team work well for you? How was the politics internally as to managing some of that, as always, that you see you laugh? The technology has its challenges that were all overcoming, but some of the internal pieces. It's tough. Our operations team struggled initially to get their head wrapped around the distributed scale out design. Finding qualified engineers who had a lot of OpenStack experience, that's also very difficult in the industry right now. So... Yeah, personnel's hard. People are learning as they go along. It's cutting in. We're all learning as we go along. We're learning from the architecture components, the containerization that's going on and everything is growing so rapidly. You either have to either embrace the technology or the technology is just going to supersede you. Yeah, we were just talking with Red Hat's SVP and one of the things we were talking about is that there's an engineering mindset going on which is fun intoxicating at the same time. It's like all the new stuff and there's some really mission critical stuff on the table that you can work on. So it's a lot of meaty things if you're into that. You build out mode, so it's appealing on one end, but a lot of the stuff's been around for a while. I mean, we've released web services, we've heard Samo on the camera that decade old. This stuff's now prime time and it's baked out. So there's some historical geekness involved around services, architectures. So does that give the balance? Does that help people walk a tight wire with a net a little bit or I mean? I think so. I mean, if we think of the first containers that were around was like Solaris zones, right? I mean, that was kind of the first containerization technology that was heavily used and adopted by user bases. But then a lot of the tools weren't ready for prime time. Now that I think we're seeing a lot of a Docker with Swarm, OpenShift tying into CloudForms for a single pane of glass and API extraction. I think we're actually seeing the birth of a new generation. You know, I know Canonical announced on Monday the new Lux decontamination and the density that they were getting from that was pretty impressive. But you know, we use OpenShift primarily and we've been very happy with it and the orchestration of how it's building into CloudForms I think is really going to unlock a lot of management and orchestration and automation that we've been trying to build. The workflow management is a huge deal. Don't you see that as an opportunity to this new generation? Absolutely, yeah. Dashboard based IT. It is, it is. And we're moving heavily away from like the traditional administrator for the most part. We're moving more towards the dev ops kind of person. Yeah, air traffic control for services, right? I mean, that's more like- Absolutely. You know, no one working the ground anymore. Everything's moving in towards an automated way. All of our deployments are automated. So when we stand up a new Cloud regardless of the availability zone or region, you know, it's all done through automation. So it's repeatable. We can guarantee the design. We can guarantee the versioning. Final question for you, Nick. What is the out in the landscape as you look forward with this, the journey you're on. You got a great couple, great accomplishment on your belt with OpenStack. You're moving forward. What do you see out there that gets you excited? I mean, what technology, you say, that's going to be a lever of change. That's certainly the next generation. What can you point to saying, I'm excited about those technologies. So I think the most excited thing I was about, you know, that I heard of today was, or actually I should say on Monday, was the Keystone integration. So Keystone right now, we have to run multiple versions of Keystone for each geolocation. Having it actually all integrated is going to be definitely a huge help for us. You know, FICO, we do a lot of security, compliance, big data analytics. So, you know, we take security very seriously. And that's going to actually help us and simplify things. Yeah, so Nick, I got the last question for you on this is, you know, your CIO, you know, gave you an edict to really go out there and be bleeding edge, you know, break some glass out there. You know, how do you give advice to, you know, people out there to say, you know, it's okay to do this. I mean, you guys work in an industry where things got to work. It's not like, you know, a little sideline things there, you know, traditional kind of enterprise workloads that, you know, must stay up and running. Why is it okay to kind of push the edge and be bleeding edge? Well, I think that everyone's kind of leaning towards that direction, you know, regardless. I mean, if you look at traditional companies like VMware, they're also embracing containerization. You know, every major company, even including Microsoft is embracing containers and the new scale design architecture. So I would say that don't be afraid, you know, embrace it. It's one of those things we're moving towards distributed architectures. There's no longer these large blocks in traditional IT. We're moving towards scalable applications. If you look at the amount of data that's being consumed in use nowadays, it's definitely moving towards a web-centric model. Nick, we really appreciate you sharing your insights and experience here on theCUBE here at OpenStack Summit. Really appreciate it. This is theCUBE bringing you all the action and all the action from the practitioners of people building OpenStack. A lot of in-production, a lot of stuff happening here on the ground. We'll be right back after the short break with more coverage. Day two of OpenStack Summit. I'm John Furrier, Stu Miniman. We'll be right back.