 Live from Vancouver, Canada, it's theCUBE. Covering OpenStack Summit North America 2018, brought to you by Red Hat, the OpenStack Foundation, and its ecosystem partners. Welcome back to theCUBE. I'm Stu Miniman here with my co-host John Troyer and we're at the OpenStack Summit 2018 here in Vancouver. One of the key topics we've been discussing, actually for a few years, but under new branding and it's really matured a bit, it's edge computing. So really happy to welcome to the program, two first time guests. We have Arturo Suarez, who's a program director with Canonical, and we also have a first time Contron employee on Eric Sorral, who's a product manager of software and services with I believe Montreal based headquarters. So thank you for allowing all of us to come up to Canada and have some fun, even if we were all working during Victoria Day, right? Yeah, yeah. All right, so let's, Arturo, we know Canonical, so we're going to talk about where you fit in. But Eric, let's start with Contron. I've got a little bit of background with them. I worked in really kind of the telco space back in the 90s, but for people that don't know Contron, maybe give us some background. Yeah, so basically the entity here today is representing the communications business unit. So what we do on that front is mostly telcos service providers. We also have a strong customer base in the media vertical. But right now at the open stack, what we're focusing on is really on edge messages as well. So we're really getting about, delivering the true story about edge because everybody has their own version of edge. Everybody has their own little precision about it down the road. It's making sure that we align every one third where it's the same messaging so that we deliver a unified solution so that everybody understands what it is. Yes, so my filter on this has been edge, it depends who you are. If you're a telecommunications vendor, when we talk to Beth Cohen, it's oh it's the edge of where they sit. If I'm an enterprise, oh it's the edge is more like the IoT devices and sometimes there's an aggregation box in between. So there's somewhere between two and four edges out there. It's like cloud, we spent a bunch of years discussing it and then we just put the term to the side and go things. When you're talking edge at Contron, what does that mean? You actually have devices. So who's your customer, what does the edge look like? So we do have customers on that front. Right now we're working with some big names out there. Basically delivering solutions for 12 inch depth racks at the bottom of radio towers or near cell sites and ultimately working our way up closer to what would look like a, what I like to call a closet data center, if you will, where we also have a platform with multiple systems that's able to be hosted in the environment. So it's really about not only having one piece of the equation but really being able to get closer to the data center. All right, and Arturo, help bring us in because we know Canonical's a software company. What's the edge mean to your customers and where does Canonical? So Canonical, we take pride of being an ubiquitous platform, right? So it doesn't matter where the edge or what the edge is, right? We do cover, there is an Ubuntu platform, there is an Ubuntu operating system for every single domain of compute going from the very end of the edge, that device that sits on your house or that drone that is flying around and you need to do some application businesses or deploy some application businesses with all the way to the core, right? Our open-sex story starts at the core but it's interesting as it goes farther from that core how the density, it's an important factor in how you do things. So we are able with Canonical to provide an operating system and tooling to tackle several of those compute domains that are part of the cloud where real estate is really expensive, right? Eric, so you all are a systems developer. Is that a fair two-word phrase? Basically we do original design. Okay, I know, yeah, okay. So I'm two steps away from hardware so that I think of those as all systems but you build things and you work with software and I think for folks that have been a little more abstract, you tend to think, well, in those towers, there must be some bespoke chips and some other stuff but nothing very sophisticated. At this point, you know, we're running or that your customers are running full open-stack installations on your system hardware in there and it's rugged and it's upgradable and can you talk a little bit about the business impact of that sort of thing as you go out and work with your customers? Certainly, so one of the challenges that we saw up there was really that from a hardware perspective, people didn't really think about making sure that once the box is shipped, how do you get the software on it, right? Typically it's a push and forget approach and this is where we saw a big gap that it doesn't make any sense for folks to figure that on their own. A lot of those people out there are actually application developers. They don't have a networking background, they don't have a hardware engineering background and the last thing they want to be doing is spending weeks, if not months, figuring out how to deploy open-stack or Kubernetes or other solutions out there. So that's where we leveraged canonical tools including Mass and Juju to really deploy that easily at scale and automated along with packaging some documentations and proper steps on out to deploy the environment quickly in a few hours instead of just sitting there scratching your head and trying to figure it out, right? Because that's the last thing they want. The minute they have the box on their end they already want to consume the resources and get up and running. So that's really the mission we want to tackle that you're not going to see from most hardware vendors out there. Yeah, it's interesting. We often talk about scale and it's our term, maybe it's a very different scale when you talk about how fast it's deployed. We're not talking about tens or hundreds of thousands of cores for one environment. It's way more distributed. Yeah, it's a different type of scale. It's still a scale but the building block is different, right? So we take the, there are orders of magnitude more of points of presence than there are data centers, right? So at that scale and the farther you go again from the core, the larger the scale it is but the building block is different and the ability to play, you know, the price of the computer is different. It goes much higher, right? So going back again, that ability to condense and open stack, the ability to deliver the Kubernetes within that little space is pretty unique, right? And while we're still figuring out what technology goes on the edge, we still need to account for, as Eric said, right? The economics of that edge pay a big part of that game, right? So there is a scale, it's in the thousands of points of presence or in the hundreds of thousands of points of presence or different buildings where you can put an edge cloud, right? So the use cases are still being defined but it's scale on a different building block. Well, Arturo, just to kind of clarify for myself, sometimes when you look at an open stack component diagram, right? There's a lot of components and I don't know how many nodes I'm going to have to run and they're all talking to each other but at the edge, even though there's powerful hardware there, there's an overhead consideration, right? Yes, absolutely, and that's going to be there and open stack might evolve but might not evolve but this is something we're tackling today, right? That's why, you know, I love the fact that the Conjun has also a Kubernetes cluster, right? That multi-technology, the real multi-cloud is a multi-technology approach to the edge, right? There are other things that we can put in the edge and, you know, the edge, as I said, it's not defined. We need to know exactly how much room you have and how you make the most out of each of your cores, each of the gigs of RAM out there. So open stack, obviously, is heavy for some parts of the edge. Conjun, with our help, has pushed that to the minimum open stack viable that allows you not to roll a track when you need to do something on that location, right? And that is as effective as it can get today. Yeah, Eric, can you help put this in kind of the framework of cloud in general? So, you know, when I think of edge, a lot of it, you know, data's going to need to go back to data centers or public cloud, multiple public cloud providers. How do your customers deal with that? Are you using Kubernetes to, you know, help them span between public cloud and the edge? So it's a mix of both. Right now we're doing some work to see how you can utilize idle processing time, along with Kubernetes scheduling and orchestration capabilities, but also open stack relicators in a more traditional SDN and NFV use case out there to run your traditional applications. So that's two things that we get out of the platform, but it's also understanding how much data you want to go back to the data center and making sure that most of the processing is as close as possible. That goes along with 5G, of course, you literally don't have the time to go back to the data center. So it's really about putting those capabilities, whether it's FPGAs, GPUs and those platform and really enabling that as close as possible to the edge or the end user, should I say. Maybe, Eric, I know you were in the carrier space. Can you talk a little, maybe a contron in general, maybe how you, in your career, as you go the next decades, right, looking at embeddable technology everywhere and kind of what do you all see as the vision of where we're headed? Wow, that's a hell of a question. That's a big question to throw on you. I think it's very interesting to see where things are going. There's a lot of cancellation and you have all these open source project that needs to work together and the fact that OpenStack is embracing the reality that Kubernetes is going to be there to drive workloads. They're not stepping on each other, so not even near. So this is where the collaboration between what we're seeing from the OpenStack Foundation along with projects from the Linux Foundation, this is really, really interesting to see this moving forward. Other projects upcoming like Onap and Acreno is going to be very interesting for in the next 24 months to see what it's going to shape into. One of the near things, you mentioned 5G and we've been watching what's available, how that rollout's going to go, the various pieces. Is this ecosystem ready for that and going to take advantage of it and how soon until it's real for customers? The hardware's ready, that's for sure. It's really going to be about making sure, if you have a split environment that's based on x86 or a split with arm, it's going to be about making sure that these environments can interact with each other. The service chaining is probably the most complicated aspect there as to what people want to be doing there. And there's a bit of rope pulling from one side to another still, but it's finally starting to put into place. I think that the fact that Acreno, which is going to bring a version of OpenStack within the Linux Foundation, this is going to be really unlocking the capabilities that are out there to deploy the solution and tying along with that, with hardware that is not, that has a single purpose, that's able to cater all the use cases and not just think about one vertical and then this box does this and this other box does another use case. I think that's the pitfall that a lot of vendors has fallen into instead of just, okay, for a second, think outside of the box, how many applications because you fit in this footprint and they're probably going to be big data and multiple use cases that are nowhere near each other, so don't try to do this very specific platform and just make sure that you're able to cater pretty much to everyone and it's probably going to do the job, right? So, yeah, there's over 40 sessions on edge computing here. Why don't you just give both of you the opportunity, give us closing remarks on the importance of edge, what you're seeing here at the show and final takeaways. Yeah, so from our side, from the canonical side again, the edge is whatever is not core that really has like different domains of compute. There is an Ubuntu for each of one of those domains. As Eric mentioned, this is important because you have a common platform, not only on the hardware perspective or the orchestrating technologies underneath which are evolving fast and we have the ability because of how we're built to accommodate or to build on all of those technologies and be able to allow developers to choose what they want to do, how they want to do it, try and try again in different types of technologies and finally get to that interesting thing, right? There is that application layer that still needs to be developed to make the best use out of the existing technologies. So it's going to be interesting to see how applications and the technologies evolve together and we are in a great position as a common platform to all of those compute domains and all of those technologies from the economical perspective. On our side, what we see, it's really about making sure that it's a density play. At the edge, the closer you go to these more wild environments, it's not data centers with 30 kilowatts per rack. You don't have the luxury of putting in what I like to call wakeboards, 36 inch servers or open compute systems. So we really want to make sure that we're able to cater to that. We do have the products for it along with the technologies that Canonical are bringing on that front. We're able to easily roll out multiple types of application for those different use cases and ultimately it's all going to be about density, power, efficiency and making sure that your time to production with the environment is as short as possible because the minute they want access to that platform you need to be ready to roll it out otherwise you're going to be lagging behind. Eric and Arturo, thanks so much for coming on the program giving us all the updates on edge computing here. For John Troyer, I'm Stu Miniman. Back with lots more coverage here from OpenSnack Summit 2018 in Vancouver. Thanks for watching theCUBE.