 Live from Seattle, Washington, it's theCUBE. Covering KubeCon and CloudNativeCon North America 2018. Brought to you by Red Hat, the CloudNative Computing Foundation and its ecosystem partners. Hey, welcome back everyone. We're live here at KubeCon CloudNativeCon. This is theCUBE's live coverage or three days of wall-to-wall coverage. Day two, I'm John Furrier with Stu Miniman. Our next guest is an end user, also a program chair of EnvoyCon, which was sold out, Matt Klein, with software engineer with Lyft. Great to have you on again. Good to see you, thanks for spending the time. Thank you, great to have you. I know you've been busy, your voice is getting hoarse. You guys had a successful EnvoyCon sold out on the front end of KubeCon and CloudNativeCon. Interesting, right? This is rising tide. What's going on, how'd that go? Why all the interest? It's been, I continue to be blown away by the overall reaction. Yeah, so we had EnvoyCon on Monday. We had, I think, almost 350 people come sold out. I think we could have had a larger room if it was available, but we didn't. And just amazing to walk around this conference and see all the cloud vendors getting behind Envoy, lots of companies building on top of Envoy, all of the end users. It just seems to be everywhere here and to have only been open source for a little over two years. I mean, it's just unbelievable. Matt, Matt, you know, I think a year ago, like service mesh was something we were still kind of, the basic understanding of what it was. And it definitely, there's certain interviews we've done this week and the like, you know, service mesh, you know, Envoy, things like Istio are going to be even bigger than Kubernetes. Yeah, well, you know, it's, I've been to the last few KubeCons and every KubeCon I think that I can't get much bigger or more nuts and no, no, everyone seems to be a little bit crazier. But no, just from the community perspective, EnvoyCon was fantastic because we had mostly end user talks. So it was really fun to get people together and to see all the different things that they're building on top of Envoy. One of the things that's, I think, impressive and I think is a real notable story, and of course we talked about it a bit last time you were on, is that Lyft as an end user kind of encapsulates and epitomizes kind of the innovation in building going on. A lot of people have been building a lot of cool stuff using Cloud Look and getting down and dirty, rolling their own and actually creating business value, not in a classic IT by IT, just build IT, build systems to build business value. And then donating it in to scale up with the community is pretty notable. So congratulations on that. Now you have startups kind of acting the same way. So the line between vendor and end user is certainly changing. I mean, we need more end users on. Well, they're all kind of end users. So this is a dynamic that is, I think, notable for this generation. And it's real. Talk about that dynamic because I think this is a real success story, but also a trend in the industry. Yeah, you know, so I think for us, what's fun for me about not only building Envoy, but seeing how it's evolved is really what you said is that I like solving actual problems for people, right? And we can have different opinions on what the different vendors are doing. Of course, there's lots of people doing different things. But for me, at least working at a company like Lyft, it's super fun to be able to build technology that solves specific problems that the business is actually happening. Now, if something becomes successful, sure, we're going to see a lot of vendors come in and hopefully build products that can help other folks. The way that I look at it, and this has been an interesting evolution for me over the last year, is I would say a year ago people would come to me and say, hey, Matt, I've heard about Envoy. I would like to use it to help solve some problems. And I went to the website and I'm like, I don't understand it, right? Like it's too complicated to use, that the documentation is not good enough. And I think over the last year, my thinking has evolved a little bit in the sense that we've seen so many people or end users or companies build fantastic products on top of Envoy. And I think one of the reasons that Envoy has become so successful is that it's a building block that other people can come and add vertical value. So whether that's a more sophisticated internet company like Lyft or a vendor or a cloud vendor, I think that's what has made the community so successful is that we can build this base thing. And it's amazing, but then we can allow people to add vertical value. And that's an interesting dynamic of both cloud and open source. You look at Amazon, one of the most successful public cloud, their core building blocks was EC2 and S3 originally. Open source is about building on top of other things. So again, the dynamic between open source and cloud scale is really kind of the magic. And just in terms of how we actually go through and I think fund some of these projects ends up being very interesting, right? So just in the sense that we have a lot of full time people working on Envoy and they're working on it actually for different reasons. We have people that are working on it as end users. We have people working on it because they're building vertical products. But in the end, everyone wins, right? Because the base technology stays technology focused. And I think that has what has been successful is that we allow people to succeed in different ways. All right, so Matt, you're at the forefront of one of the most difficult problems that we're looking at these days. It's scale, distributed systems and edge and how that ties in. I want to just get your kind of macro level viewpoint as to how we're doing this industry. What are some of those tough challenges we talk about? We talk about things like IoT and Edge and vehicles of course have a lot of them. Yeah, so I mean, I think when you say scale there's two things that comes to mind. There's physical scale and I do agree actually that we are continuing to push more compute out to the edge. And in fact, I talked about this a little at EnvoyCon but I have I think some very exciting projects to bring or plans to bring Envoy actually to mobile phones and to edge devices starting next year. So all more to say about that in the spring. I'm very excited about that. So I do think that there's a lot of opportunity to better evolve how we ingress data from the edge, how we do compute out the edge, a bunch of other things and I think Envoy will be at the forefront of that. But when you talk about scale, I still think that there's a lot of human scale involved of how we scale the number of developers that are working on all of these architectures. And I do think that service mesh and Kubernetes and a bunch of other stuff, ultimately if we're successful, it helps us grow the number of product developers that can successfully work on these systems. I still think we have a long way to go. But I think that's one of those areas where I think some of these technologies help people both at physical internet scale but also at human scale. Well, I really appreciate your work that you're doing. You can contribute to the community both on solving the problems with Envoy and also being the program chair of EnvoyCon. I think it's going to be great for the community. I got to ask you, as you get pulled into a lot of these, I won't say political or media kind of conversations, you got to kind of be a helicopter and get above and get high level and talk to people who are discovering and learning for the first time, which is part of what communities do. How do you talk about those other end users that's saying, hey Matt, our companies are going to reshape their IT investments all based on open source. And I really want to learn more about Envoy and just the benefits of cloud native in general. I got to go, and I believe her, I got to go talk to some one of believers or non-believers in my company. And I got to make my point home. How do they be successful? What's your advice to that? Because that's a challenge a lot of people are having. I totally agree. My advice first and foremost is to start by understanding what problems are trying to be solved. And I actually think that sounds very obvious but I think that people don't do it enough because I think sometimes we come to conferences like this and we see all the amazing technology that people are building and it seems fantastic. But if one tries to adopt everything that they see here without understanding the incremental steps and the things that are the problems that are being solved, that can be very problematic. It's a new kind of technical depth. It's kind of a new way to... So my advice is to start with what are the actual problems, right? And whether that be observability issues or authentication issues or security issues or whatever is to start with the problems and then work backwards. And my advice is always incremental, no big bang, right? And try to figure out the right incremental path of adopting the smallest piece of technology that solves a particular problem and go from there. And build economies of scale to the mission. Right, and whether that means working with a vendor or working with the raw open source technology, that's a personal decision of each company to figure out what their comfort level is. But that really is my advice, is start with the problem statement and then figure out the easiest and the quickest incremental path forward. The trends that we're seeing, Stu was talking earlier, a lot of hyperscalers near, a lot of diversity coming into the community, just what's the hallway conversation amongst the people in the community around as the community grows larger? I mean, open source community core persona for constituency, then you got the downstream impact to that is IT is changing, developers are coming in. So it's not so much changing personas and target audiences of the environment, open source is still core, that's kind of the downstream impacts. And so you see a lot of people come in, IT people, new developers, how does the community look at that? What's your view on how to engage but also not alienate new people? Well, I mean, I think ultimately, we are attempting to build systems that help people be successful and be more productive, right? And I think the natural evolution of that is bringing some of this technology into the enterprise. We have to recognize that as the community scales, the baseline level of knowledge is different. I mean, we all come at it with different understanding of whether it be networking or orchestration or security. And I think what I would say is that we're never going to build one technology that makes everyone happy, it is impossible. It's impossible to build a technology that satisfies both the expert user and the entry level user. So I believe that we need to build layered technologies, layered abstraction that allow people to plug in at different levels and some of them are more opinionated than others. And I think it is recognizing and supporting a community that has base level technology, has vendors adding value at different layers to help people and really just respecting the fact that people come at it with different levels. The application assembly is really the where it's going. Exactly, I agree. Matt, I'm wondering if you could reflect back for us. You're the creator of Envoy. I saw you up on stage yesterday, just the supportive team and the community that helped us grow and you've reached graduation. What does that mean to you, to the team? Because it's different than a school graduation. This is not the end of something, you don't get a diploma at it. Is there a party? I don't know if there was, I don't think they invited me. The pictures. The foundation is picking up the martyrdom. I don't know. Maybe. So, from a project perspective in terms of how we go about our day to day, I don't think that much changes. I think we have been operating as a mature, graduated level project, probably for quite some time in terms of adoption and methodology and stuff like that. I think what graduation means for the project is it's a vote of respect from the larger industry and the community that Envoy isn't going to disappear. It's not going to become an abandoned project on GitHub if, for example, Lyft stops investing in it. I think we've reached a critical mass of project success. And I think what that means is that it allows folks that may be at more conservative organizations, like who may be a little later to adopt newer technologies to give them the confidence that says that Envoy is not going to disappear, that we can potentially bet some of our future on Envoy. So, I think it's a vote of confidence. I don't think it changes a lot about how we operate on a day-to-day basis. Matt, thanks for coming on theCUBE. And again, congratulations, seminal work. You guys are doing great, and Lyft is really, I think a great example of the new dynamic and open source where they're building and they're working with the community to continue to extend that. And this is what we want. That's what open source is all about. It is. So, congratulations. And we have a graduation party for Envoy. We'll figure it out, we'll put photos and pictures and everything else. Thanks for coming on theCUBE. Thank you very much. I appreciate it. theCUBE coverage here live. I'm John Furrier with Stu Miniman. More coverage after this short break. Stay with us.