 Live from the Computer History Museum in Mountain View, California. It's theCUBE, covering DevNet Create 2018. Brought to you by Cisco. Okay, welcome back everyone. We're live here in Mountain View, California, heart of Silicon Valley for Cisco's DevNet Create. This is their new developer outreach, kind of a cloud DevOps conference, different from DevNet, their core Cisco networking developer conferences, kind of an extension, kind of forging new ground. Of course theCUBE's covering, we love DevOps, we love Cloud, I'm John Furrier with Lauren Cooney, my co-host today, our next guest is Alex Ellis, Project Founder of OpenFast, F-A-A-S, Function as a Service, that's Serverless, that's Kubernetes, that's Container Madness, you name it, that's the cool, important trend. Thanks for joining us. Yeah, thanks for having me, it's great to be here. So talk about the founding of the project. So you're the founder of the project, and you now work for VMware, so let's just get this on the record. So, take a minute to explain. This is important just to set a bit of context now. I started this project from the lens of working with AWS Lambda as a Docker captain. I was writing these Alexa skills, and I found that I had to hack in a web editor and click upload, or I had to write a zip file, put dependencies on my laptop, and upload that to the cloud every time I changed it. It just didn't feel right, because I was so bought into containers, it's the same everywhere, there's no more, it works on my machine. Right? So I put a POC together for Docker Swarm, and nobody had done it at that point, and it got really popular. I got to DockerCon called Hacks Contest, and presented to 4,000 people in the closing keynote. And I kind of thought it would just blossom overnight, it would explode, but it didn't happen. And actually, the months we're going back 14 now, I grew a community and spent most of my time growing the community and extending the project. Now that has been really fruitful, it's led to over 11,000 stars on GitHub, 91 individual contributors, and much, much more. It's been a really rich experience, but at the same time. It's a rather big rocket ship, you kind of went hunkered down and got a kernel of core people together. Yeah. Kind of set the DNA, what is the DNA of this project? You had to describe it. Yeah, so I think at the heart of it, it's serverless functions made simple for Docker and Kubernetes. Great, and so how does Amazon play into this? You were using Amazon Cloud? I was using AWS and I was using Lambda and that flow was not what I was used in the enterprise. It wasn't what I was used to as a Docker captain. You know, I wanted a finite image that I could scan for vulnerabilities, I could check off and promote through an environment. Couldn't do it, so that was what OpenFaz aimed to do was to make those serverless functions easy with Docker as a runtime. Well, congratulations, it's a lot of hard work. First of all, building a community is very difficult and certainly one that's relevant. Cool and relevant, I would say, is serverless and functions. Certainly we're seeing that now at the uptake. Still early on, but people are working on it. So then now, forward to today, you work for VMware. So how did they get involved? Are you shipping the project to VMware? Do they own it? Do you maintain the independence? What's the relationship between VMware yourself and the project if you can talk about that? Yeah, I think that's a great question. So I got to the point where I had demands of my time around the clock. I couldn't rest, open source project, weekends, nights, a lot. You need beer money too, by the way. You need beer money. And I was working at ADP and just doing all of this in my own time and had a number of different options that came up and people saying, look, how are you going to sustain this? How are you going to keep doing what you love? You should be working on it full time. One of the options that came up was from VMware to work in the open source technology center. It's relatively new. And the mission of the OSTC is to show VMware as a good citizen in the community and to contribute back to meaningful projects that relate to their products. Yeah, and they have good leadership there too at VMware. A lot of people don't know that. We did a couple of CUBE interviews with them last year. And there is a group inside VMware that just does that, not with the tentacles of VMware and Dell Technologies in there. It's an independent group. They probably go to some meetings and do some debrief. But the most part is kind of decoupled from VMware, right? So the mission is not necessarily to make money and to produce products. It's to contribute to open source, help with inbound. So when we need to consume a project in a product and outbound when we want to make the world a better place. So I'm not going to put words in VMware's mouth, but I think I will speculate covering VMware since the CUBE started. We've been to every VMware world and everyone knows we've got the good presence there. But if I'm VMware, I'm like, hey, you know what? We just did a deal with Amazon. Our enterprise group is not so cloud savvy. I mean, I was the enterprise, they're operators, not true cloud native, but they're bridging that gap. The world of cloud native and enterprises coming together. Does this project fit into that spot? Is that kind of where they saw it? Did I get that right or what was their interest other than helping the world out and solving world peace in the open source community? Yeah, so the mission of OSTC is slightly different. It's to contribute back to meaningful projects and to have this presence in the community. Now, I think OpenVas is particularly attractive because it has such a broad community. There's people all around the world that are contributing to it, very active. For VMware, it makes a lot of sense because it runs natively on Kubernetes or Docker Swarm and it's gained a lot of traction. People are using it. I had a call with BT Research before I came out and they said we've been using it for seven months. We absolutely love it. It's transforming how we're doing on microservices. And so I think that's part of it as well because we already have kind of a lead. We already have a lot of momentum with this project. So are you looking to, I know that the organization that you work for is really focused on driving this outbound, right? So, but is VMware using this internally as well? So I think there's been a number of people who've shown an interest. You can think, right, there's a problem we could solve with this. And I'm just getting my feet under the table, but really my mission is to make serverless functions simple to build this community and to have something that people can turn to as an alternative. So one of the things that I did in the talk yesterday was how do you explain OpenFast to your boss? And one of the points there was to unlock your data. And I think we talked about this briefly before. Now with controversies recently about data and who owns it, what's happening with it, I think it's even more relevant that you couldn't have full control over the whole stack if you want or use a product like Microsoft AKS, their Kubernetes service or GKE and actually treat OpenFast like a very thin layer of automation or go full stack and have everything under your control. Yeah, I mean, that's a great conversation I have too because obviously you're kind of referring to the Facebook situation. Zuckerberg's testifying for a Senate yesterday, Congress today, and it's funny because watching him talk to senators in the US, they really don't know how stuff works. And so if you think about what Facebook does, I mean, granted, they took some liberties. They're not the perfect citizen. They got slapped. They took it through the woodshed, if you will. But their mission is to use the data. And this is where cloud native is interesting and I think I want to get your reaction to this. You need to use the data, not treat it as a siloed, fenced in data warehouse. That model's old, right? It's now horizontally scalable. Data's got to move and you've got to have data to make other things happen. That's the way these services are working. So it's really important to have addressability to the data and GDPR takes an attempt at kind of hand-waving that simple argument away. I'm not really a big fan of that personally, but the role of data is super important. You got to make it pervasive. So the challenge is you manage those controls. Is that an opportunity for functions? What's your reaction to that whole paradigm of data? Yeah, so we're talking about anonymous usage data like with the Facebook situation or... Just data in general, if I'm an application and I have data that I'm generating the same derivative service, I need, you might want to leverage that data. So I'm going to have to have a mechanism to share that data, to make your service better because data makes data on the side of it is interesting. But then if you get trapped in regulation and licensing, it can be disruptive. So as an engineer and an open source engineer, you find people that have no clue about what an MIT license is to a GPL or why you'd use one or the other. I think there's a lot we can do to educate the wider community and help them understand the basics of these issues. When I was at university, we had a course on ethics and legal issues and licensing. And I heard on the radio earlier on the Uber that they're starting to try and up the level of that again. And I think it really needs to start at a ground level. We need to educate people about these issues so that they're aware of how to handle the data. I mean, if you look at common tools like Docker and VS Code and Atom, popular editors, they collect anonymous usage statistics and you have to opt out. Should open files collect data as well? Because it could be super helpful for us to know the right thing to do. And when you come to open source, you get no feedback until somebody wants support from you and it has to be done yesterday for free. And so yeah, getting data could be super powerful. Wow, she bring up a great point. I think this is something that's worthy of an ongoing conversation. I think it will be too because GPL, Apache license, all these licenses were built when open source was a tier two citizen. So the whole idea of these licenses was to create a robust sharing economy of code and you know, with the certain nuances of those licenses. But just like stacks get updated and modernized with what we've seen with containers and now Kubernetes and serverless, the stack is changing and modernizing. The licenses have to do it as well. So I think this is something that, I think it's kind of like, we got to get on it. I think we just, this is a work area. It's not necessarily, it's game changing if you don't do it, right? Because it could flip it either way. To me, that's my opinion. Well, I think you're under MIT, correct? So it's under MIT right now. One of the things that I didn't realize when I started the project is if you want to get into a big foundation like the Cloud Native Computing Foundation, you need an Apache 2.0 license. And the main difference is that it offers some protections around patent claims, but it's basically compatible. So it is a minefield. So that's just for the CNCF. Right, and the Apache foundation obviously as well. And probably many others follow suit because I think it's, we talk about this new stack. It's a dual source issue. A refresh, right? It's a compatible license. It seems to help a lot of people. That's a huge issue because you could have been well down the road with the important code and then the lawyers will make you take it out. Right, so that's why organizations like the Open Source Program Office exist within VMware to help these issues and to monitor and do compliance. They may use software like BlackDuck to check stuff automatically because you don't want to be doing checks on your aircraft once it's in the air. You want to sort everything out on the ground. You'd be grounded your fleet, that's for sure. When it comes to that. How do you handle that with that licensing? How do you guys handle that when you talk when people contribute? Are they aware of the license? Are they understanding the implications? So with Open FAS, we're following a model very similar to the Linux kernel, which is a sign off developer certificate of origin. What you're saying is I'm allowed to give you this code. I'm allowed to be a part of the project. And I wrote it, I originated it. And that's pretty much a good balance between a full contributor license agreement and nothing at all. But look, there's a lot of projects in this space right now. I don't know if you've noticed that. Kubernetes serverless projects. Yeah, I mean, it's a lot of, it's why I like this show here. I think what Cisco is smart to do with DevNet Create is identify the network programmability which really takes DevOps, expands the aperture of what DevOps is. So as you got new applications coming online, some developers want nothing to do with the infrastructure. Kubernetes has got a much more active and more prominent role with layer seven primitives, for instance, or managing things down to the network layer. You're talking about policy services, inside services on the fly. So this is really a good thing in my opinion. So I think Kubernetes, most people look at it as a kind of generic orchestration, but I think there's so much more there. I think that to me is attracting some really rockstar developers. Yeah, well I think the fact that you are open, you're under the MIT license, which I'm a fan of. And it is, you're on a very successful trajectory in terms of what you're building and who's engaged and the fact that VMWare is behind you means that they're going to put some money into it, hopefully, and help you guys along as it works. But it is also a project that is not, it doesn't have folks just from VMWare. It's really, really diverse in terms of who's committing the code. So I think there's a lot of things that are really going for you. Now who do you see, you mentioned competitors. So can you talk a little bit about what the ecosystem there looks like? Yeah, so there's a number of projects that I think have made some really good decisions about their architecture and their implementation. But they all vary quite subtly, and one of the questions I get asked a lot is, how is this different from X, Q plus, Nucleo? And if you look at the CNCF landscape, there used to be a very small section with open files, Lambda and a couple of others. It's now so big it has its own PDF, just about serverless. And I think that's super confusing for people. So part of what we're trying to do is make that simple and say, look, there may be many options. Here's open files, here's how it works. You can get it deployed in 60 seconds. You can have any binary or any programming language you want. And it will scale up over Kubernetes. We'll just make a really deep integration, give you everything you'd expect, really nice developer experience. That's great. What's some of the use cases you guys see right now, low-hanging fruit for developers that want to come in and get involved in the project? Have you guys identified any low-hanging fruit use cases? So what I've seen, and I talked about this a bit yesterday in the talk, is three big use cases, really. The first one was Anisha Kashavan at the University of Washington. Now she's doing a lot of data science with neuroinformatics, medical images. She's able to take scans of brains and give them to people like you and me who don't know anything about medical science. We just draw around the lesions and we train her model. And then she makes it competitive like a game, gamifies it, you get more points. But actually what we're doing is making the world a better place by training her medical imaging database. She'll then use that as an open files function to test real images as part of her post-doctorate. So she's crowdsourcing, wisdom of crowds, collective intelligence for her research. Now one of the other things that I think is really cool is in the community, we built out a project with two 17-year-olds. Two 17-year-olds built a really cool project. Now when I think back to when I was 15, 16, I was playing with something like PHP on Windows, Lampstack, you know, I had to do everything myself. They got this scaffolding built up and they could just go to the 10th story and just keep adding on. And they didn't have to worry about managing this infrastructure at all. Or architecture, foundation of architecture. Yeah, it's gone, and that's exactly the reason why you want to do that. So they wrote some small blocks of Python that we found this machine learning code that could convert a black and white image to color, wrapped it in a box and said there's a function, then dropped it into open files and started feeding tweets in. That was pretty much it. Now we have act colorizebot. There's a bit of a strange spelling but you'll find it on Twitter and it's been in LeMond newspaper all around the world. It was pronounced at KubeCon as well. And it's just a super interesting way of showing how you can take something very complex, right? And democratize it. Yeah, we'd love to get those people working for the Kube but with KubeBox and throwing all the tweets in there. Alex, thanks for coming on, congratulations. What's next in your project? Tell us what's going on. What's next for you? What are you guys conquering next? So I'm really focused on growing the team and community. We've got an open recruitment position open right now and a small team that's building internally. I think the more people we can get, contributing on a regular basis, the more support there's gonna be for the community, the more people are gonna want to use this. We actually had 26 people join a call last week how to contribute to open files, that was the name of it, around the world. And the best part for me was where we got to the testimonies and I had people just showing their tips and experiences of how rewarding it is to contribute something bigger, something that you as a developer actually want to use. Yeah, and the value opportunities to extract value out of the group is phenomenal. Functions as a service, super relevant in cloud and DevOps as the middleware, if you want to call it, that expands. More capabilities in DevOps are coming. It's the Kube coverage here at DevNet Create. We'll be back with more live coverage here in Silicon Valley and Mountain View, California after this short break.