 Alrighty, then I'm gonna get started guys in the back there. Okay so I Titled this after a blog post I wrote about four months ago About trying not to reinvent the wheel at at OpenShift and I submitted the topic Sort of as a poke at OpenStack to get them thinking about you Cross-community collaboration and the effect that it could have on OpenStack I didn't expect them to accept the paper. I thought they would like say absolutely not We have to incubate everything. So I'm gonna talk a little bit. My name is Diane Mueller. I work at Red Hat I am the OpenShift community manager. I am a bad Python programmer I was in love with Django when Twitter came out. So I Conclamorated the two and came up with my hash What was my name on IRC for a long time and that and I used to be a DJ and I'm a bit of a snake charmer Which makes me a very good Community manager and so what I'm going to try and do today is tell a few stories about What we're doing at OpenShift what I've learned at Red Hat over the years and Share them with you and see if there's something you can take away And talk a little bit about my experience working with the OpenStack community as well And then talk a little bit about the road ahead if that makes sense. So at Red Hat There are over a hundred thousand open source projects that they sponsor or they have resources working on and a number of them are Represented in this graphic and a lot of them are in the booth here this week, too And they all have to work together. We have to collaborate internally So we do a lot of internal collaborations in the project that I work on is OpenShift origin Which is a platform as a service that runs on infrastructure as a service, which is what OpenStack is it runs on Amazon It runs on bare metal. It runs on cloud stack But we also work in on that platform. We work with lots of other communities that provide Technologies frameworks languages that run within are and on the platform as a service So there's a lot of communities that we have to talk to Interact with make sure we have the latest versions of their software and everything plays well together So there's a lot of magic in keeping all of that working together But I'm gonna tell you a little bit about my background is I started a long time ago My first job out of university was working for a small company called Nike in Hillsborough, Oregon And my job at the time was designing the the software and Using the software that designed the bottoms of your sneakers back in the early 80s So it was the feeds and speeds and all the math and stuff like that and every bit of that software It was the same software that was being used at McDonald Douglas aircraft is design aircraft parts. So when I went for training, I would train with the same guys that were They were designing airplane engines, and I was designing sneaker bottoms It was slightly a different world, but everything about that was proprietary And then a revolution came how many of you have read the cathedral and bazaar if you haven't you should It really changed the way that I looked at and the world looked at software So open source became sort of a revolution and then another revolution Happened with all that software being available and software started being the thing that was driving Organizations and giving us the ability to bring new services and new software into into the world for us And then cloud came and it changed everything again and the combination of clouds Computing and open stack the open source stuff has really changed the way My world looks because now the same software that I was using 30 years ago to dry to do Sneaker bottoms and create those molds I now have 3d printers at home can that I can do the exact same thing on and a lot of that is based on Open source software as well so Everything completely changed in the way that I looked at software from where I started in the proprietary world Because of two I say major shifts opens open source and cloud computing The other thing I tell a lot of stories. So how many of you have heard the story of stone soup? All right, it's an old one. My dad you my dad is Germany and He used to read this book. This is exact book to us when we were kids It's a story of a soldier that comes to a town and he has a rucksack on his back and he has one rock wrapped in paper and he brings it to the town commons and he Coverses the the folks there to build a fire and put a kettle of water on and he says I can make this magic soup for you I will put this stone into the water and it will become the most amazing soup in the world and This the story goes on and he Farmer walks by has a handful of carrots and he talks to they says well It would be so much better with a few carrots in it and a few in that the Butcher walks by with a ham bone and eventually you get the picture he courses and works through I'm getting all of these people to add to his project the soup and in the end He has a bowl of wonderful soup and he shares all that soup with everybody And they that's the lesson that I grew up with and that's kind of how I think of community is you it's a lot about Communication some of its coercion, but it's all about doing something that benefits everybody in the end But someone is driving it someone is moving it forward and someone has to instigate it and bring something to the table to get it started and kicked off The other book that I really love and how many of you have read mythical man month So this sort of dates me I think it came out over 25 years ago and it really talks about as Projects grow in size and as we're experienced here I think we're almost up to 5,000 attendees here at open stack summit the number of interfaces between all those people becomes so much more complex and The the key to success to any project is really the communication that we're able to bring and the facilitation of that Communication and open stack has done a lot of great work about creating an amazing community That collaborates together very very well But adding more people to that project isn't always the way to fix things So having 5,000 people here is great and it's a wonderful thing and it's a great show of market size and the growth of the community But it doesn't always solve all of our problems So I'm going to tell a few stories about what's going on on the project that I work on And that is OpenShift has a lot of communities that we have to touch base with we have an online community of Over two million application developers who deployed on our publicly hosted software As a platform as a service at OpenShift online So we have this community of developers that really we only touch in that they deploy their applications on our Platform on our cloud and then we have a huge community of contributors. There's a few of them in the room Here today that are working with us to develop that code base and then we have a larger installed base of Enterprise customers people like Telstra and Orange and a number of others that are really been amazing people and All of this feedback keeps coming into the project, but it comes from different use cases and different stories So one, you know, and we've been a mild success I would say some there's a couple of surveys out there and everybody can find a different survey It says wonderful things about themselves But one game out where really we've really made a good impact on the private pause landscape So we have had some modicum of success and the super user survey that came out today Said about 11% of you who are using Application configuration tools are using OpenShift. So we've been moderately successful We don't brag a whole lot But for us, it's really never been just about the contributors to the code base We I think when I took this screenshot about a month ago We had about 190 contributors and other good I'd say a good majority of them come from red hat engineering and a number of them come from outside But we've got a steady growth pattern going on With people who are contributing code And we're really been very good in the Octoverse last year We came in on which was a huge surprise to us We didn't we weren't really tracking that but of the merged pull requests in on github We were the fifth in place of people who would take in your code and merge it into our user base So we you know done some really cool stuff Cisco's done some amazing Contributions and other companies have to so you know the code we're moderately successful in getting Contribution to our code base But it's never just been about that our real success has become has been in sharing the knowledge across The different constituencies and the experience and the best practices And the ideas about how OpenShift could be used in different contexts in different business models Not just by developers, but by hosting providers and operators So when we look at Community we don't mean just the developers who are working on the core code I kind of think of it as as a community manager as I have an Olympic-sized task. I have to Go across all the related open-source projects like this one here open stack the work we do with in It with Ruby and Rails and v2 with the go community with all of the Docker and Kubernetes folks that we're working in v3 all of our wide variety of service providers because in that layer cake We have all kinds of cloud era and Hortonworks and lots of different people who we have to connect with and Bring their services into up into the fold And we have a lots of wonderful partners that we've been doing with So something that we've done for kind of quietly on the side is we've been working on creating this concept sort of like wiki commons or flicker commons if you've used those things that Allows us to create a peer-to-peer network Of people who are using contributing to the different open-source projects that we've been working on And this has been something that we're pretty proud of that. We've gotten a lot of people to connect And in some ways it's about creating those lines of communication the virtual Meeting spaces special interest interest groups and mailing lists where kind of red hat can just sort of step out of the way And let the the participants have those conversations and we do things like set up briefings for Kubernetes people to talk to hosting providers and and open stack people So we've got a pretty extensive list of folks that have been our mix of customers and open-source projects Hosting providers and so we have some really wonderful conversations in the commons And you can find out more about that if you'd like to join or or learn more at the commons one So for us, it's really been focused on scaling out the connections within the community so In order to make Open-shift a successful community. We've had to figure out ways to do that. We've done we've taken on something Sort of a copy of Fedora ambassadors program We've created something called the open-shift accelerator program, which I'm hoping we have about 40 or 50 members of that and basically they're cloning Myself and other people anyone who goes out into the world who presents or wants to talk about open-shift We give them the tools because we in growing the community of the size We have to have those kinds of ambassadors going out. We can't be everywhere Though I think I feel like I have been the past couple of months traveling pretty frequently I'd like to be able to say, you know Someone at Cisco can go and give this presentation or someone at Telstra can do that or someone else Is perfectly capable and so we've been sharing that information too and growing that community because for us, it's really about making sure that people have all the tools and Pieces of information so that we can connect help them connect the dots with each other and then get out of the way Because community really is about that connection and that collaboration. It's not about being force-fed anything So the reason I titled this Presentation was one of the things that that has also helped us be a real success I think at an open-shift and a lesson that we've we've learned probably the hard way Is that we need to collaborate with other open-source? projects, so we need to be able to take software co-basis functionality that are not our core competencies and and bring them into the fold and utilize them successfully and One of the things that brought me in to start back in back in Essex in Boston I started talking With open stack and then when heat came around one of the things we did was we were Really closely with the heat folks and created templates Some of the very first heat templates for testing out deploying an auto-scaling open shift on on open stack And that gave us a really good understanding of how to make everything work how to spin up Nova Instances how to register something with glance we learned a lot from that, but they also in turn The heat team learned a lot from taking a complex application like Open-shift, which is got brokers and nodes and has to spin those things up and make sure everything scales They learned a lot from it too So that collaboration was really helpful for us to make sure that we played well with the tooling that the open-stack Community wanted us to use the local orchestration stuff And we didn't do any reinvention, but we both learned from each other And so that's sort of where we're at in terms of the open-stack community is really making sure that not just within red hat But outside of red hat we start working with other open-source communities So we don't create tooling that's unnecessary and we use the tooling we we also have if you go to install that Open-shift comm we've done a lot of work with the puppet community So we have we try and use the tools that exist In your enterprises and in your development shops today rather than reinventing the wheel The other tactic that we use a lot and we've more frequently this year. We've been doing is a what I call Embedding resources, it's like taking someone right off the open-shift team and putting them into somebody else's Open-source project and this works really well Oh It's been shift v3 if you didn't see the presentation the other day. It's videotape, but it's a complete rewrite of Open-shift in a new language go So we were Ruby in Rails we rewrote it and go Which is great if you didn't like installing Ruby and creating Ruby gems on your desktop for the CLI to work That's gone now But we didn't we ended up using a new language Everybody's favorite topic this week apparently is is Docker. We took a lot of red hat resources from the rel team and from the Open-shift team and Embedded them and put them right into the Docker community to work on things like bringing se linux and security Into Docker and working to make sure that we understood where Docker was going and how it was working So that we could then use Docker containers in v3 successfully So this trade-off between different groups has really Gotten a lot of helped us a lot move to the next generation Had we not have we continued down the path of using and baking our own Containerization approach we probably would be a very stale product at the moment the other thing that we've done and We're continuing to do is to work very hard with with the Kubernetes community as well And I think we have Clayton Coleman from the Open-shift team has now got commit rights and has is probably one of the most Probably the hardest working person I've ever met but has done tons of Commits and into the Kubernetes product project and they've been very good about bringing that and allowing us to work with them and learn from their Production scale capabilities to bring that into the Open-shift world. So again, it's like Google has open sourced an amazing project They've they've learned they're sort of opening up the kimono on everything that they know about scaling containers and We are being able to take that in and move our our offering to the next level of scaling Capabilities and it's been pretty a pretty amazing ride and it continues to be a project atomic is a red hat sponsored project that We're using pretty heavily as well inside of Open-Shift So we're learning about getting that lightweight level of OS available too. So there's we do a lot of work with them There's a SIG that we have now. That's a special interest group That's with rel atomic and Open-Shift if you're interested in that those three groups are all helping each other to use that and embed it We do incubate some things And incubation is is a tricky subject and it can be kind of touchy sometimes We we have a small project called gear D that is meant to help if you have a docker container help you wire it up So it can be aware and build the docker containers The trajectory of gear D is I'm betting That it will be consumed and that's functionality of it will end up actually being consumed into another project Maybe into Kubernetes itself at some point. So it's not that we're against Incubating stuff within the open source because you the gear D stuff is in the Open-Shift repo And there's a few other small Open-Shift extras that are in there. So when we need to we will Build new functionality in and but what our our ethos is more of trying to figure out how we can do that Not that we're trying to offload the work into other communities because we'll put the resources in there to make sure That it's maintained and supported but really where it's appropriate because this is What allows us I think to iterate quickly to to rearchitect and to take advantage of new Technologies as they come out and so the the platform as a service Can that offering we have can move forward? Architecturally rather rapidly because we didn't wire it all up ourselves. Everything's pretty modular the apis are pretty Good so that we can take advantage of other technologies and we don't have to invent all of it here so There's some trouble with incubation Because we really get to because we do this We do embed some resources so we do lose some resource time into these other projects But allows our core engineers to really focus on the value add for our offering And for our project so we're not distracted by trying to figure out how to do Orchestration at huge scale we take advantage of Google stuff And if you look a little bit at what OpenStack has done We've got some amazing projects here and Red Hat has people probably on every single one of these projects But when it gets pretty scary at times if you start to think about all the different pieces and parts of OpenStack and I've been you know playing with it for since the days of Essex for for quite a while And but it is it is getting better It's definitely I think people are starting to look and yesterday. We had a couple of sessions that were pretty good starting to think about how we can really In turn change the focus to a more layered approach And so that we can make sure that at least the lower the base computer the base layer is Interoperable in all the distributions that it is that we make sure we get all of the different pieces apart This is a photo from yesterday's two sessions. We they did do back-to-back sessions on growth yesterday in the design summit And you know there was a some debate about the big tent approach You know whether we want to have this as OpenStack this huge big big tent and continued to incubate and grow new things And there's a sort of an ethos in an OpenStack of Inclusivity and I think that's because our culture is tries to be very inclusive and tries to promote diversity but in some ways the some of the scaling issues of of resources and People who have core privileges and who are doing reviews on the pieces of that very base layer are lagging behind and we're not getting the core pieces or the base layer Focused on by resources the way that we could because we're distracted by other projects in my humble opinion So a funny thing happened on the way to the foundation I Stole this picture from Randy Bias's presentation at Silicon Valley because He did a great presentation about Maybe not the need for a benevolent dictator for OpenStack, but the need for product management and I when we get into Creating the foundation if you talk to everybody on the different parts of the foundation and everyone has a perspective on What OpenStack is and I think one of the things? if you could Take that base layer and if we could focus our resources on making sure that that base layer was solid and that was interoperable We would have a better time of getting people to adopt it into getting more OpenStack out there in the world and then the ancillary pieces that we are currently setting up Maybe have could have better homes elsewhere They would be third-party projects or maybe they'd be in the Apache Foundation as opposed to OpenStack Foundation projects and I think that's one of the things that happens when you work on industry-driven Foundations or when there's a lot of people who are thinking about it from That from where they're there where their commercial value is going to come from as opposed to something that's market-driven and this again I think is comes from maybe a lack of product management in The oversight of the project and it is getting addressed But it is it's a huge differentiator from other projects. So like Linux was much more Driven by the needs and the use cases and and where people Really had needs to work on things as opposed to trying to add on additional projects So with OpenShift with OpenShift with OpenStack, it's already sailed. We've already created the foundation. It is industry-driven Driven that's pretty much the way that we what we have to live with and I think we're doing a pretty good job of responding to that and Working together hopefully to get some more product management. It's there but it what it does And my my opinion I'm talking on on my opinion and not red hats I'm gonna take is that by having that industry driven point of view We really risk of creating a new cathedral. So instead of something or a lot of new cathedrals Maybe that that create a lot of risk for us in terms of having a project that people can easily deploy and And scale out and use in lots of different instances And this has been a conversation. That's been pretty been running through a lot of the threads in in the in this conference in prior conferences, too Is that how do we make sure that we don't lose our focus on that base project? so they're in my Experience and I've been around for almost 30 years is there's sort of two ways that foundations really work is and it's and they're Basically and both of them service is a key piece of it like I've been Working with Python for many years and it and it's in service to attack that the technology that Python is and they really do an amazing job with Making language distributions and in vetting new modules and things and then there's the infrastructure Providing infrastructure as a service basically for projects that Apache does and we sort of have to decide I think from an open stack Perspective which we are in or how we're going to blend those two and make sure that works for us And it is a you know all about service And it there might be a place in the Apache foundation for some of the things that we're trying to ink We think we need to incubate with an open stack Another thing a lot of the times most of the other foundations if you look at the history of Mozilla and Linux is that The foundations came along after the projects had some maturity. There was people wanted to make sure a stable Well-maintained code base happened and continued to be available. So I think Historically when choosing for to do a foundation or something you really have to be Cognizant of what the status of your project is at that point in time so I'm sort of getting ahead of myself and talking probably too fast, but There's a couple of the lessons that we've learned really is to very carefully incubate And learn just to say no to new projects. Don't try and create new tooling when there's good tooling out there already Talk to your user base what they want. We have in our case. We use puppet We have ansible installers. We've we've been working with the heat community And we're making sure that we're using the tooling and we're not reinventing tooling that Other people have done a better job and have more resources to work on and collaborate and embed wherever possible So that's easier to make and make sure that your software layer that you're creating can work with other projects its key is to maintain the culture of your project so the the inclusivity of being able to include other Partners and users and folks in your project is pretty key Being able to make sure that you have really strong communication pathways And this is not just the one-way pathway from red hat out to the users and the users back into red hat This is one of the major reasons we created this commons initiative was so that we could make sure that our user base could talk amongst themselves and Talk to each other and that was that's been really a key Factor I think and the success is that we don't Try and stand in the way and we make sure that everybody has a I think open stack has done a really good job of growing the leadership and and communicating the vision of the culture and that but I think there's some work to be done to focus what is open stack and to be clarify that and Really the key I think is to really start to be able to figure out how to connect the dots between all of the different communities So we see in the marketplace. We see all the vendors that come here and promote their distribution or their Additional things, but what we need to do is get more of the users talking to each other and build Maybe an open stack commons so that we can get that peer-to-peer Conversation going and get the operators and the users Talking more to each other and building those communities out there in those pathways and that really is I think one of the keys to success is to really try not to incubate Too many projects to make sure the communication between not just all of the core developers But through all of the community members and that they all have access to each other And that's one of the success factors that I've seen So I this is my stone. This is the soup that I'm trying to to feed you all I'm not didn't I don't expect everyone to agree with everything. I've said here today But this is some of what we've learned at OpenShift and that has allowed us to really easily Rapidly iterate into the next generation of code and when we see the weight of having some of these extra Projects and baggage that the open-stack community has to bring along with it We're hoping that we can share some of these lessons and start to get some more cross-community Collaboration going within the open-stack world And that's That's what I had for today. So I'll take any questions anyone might have but thank you. Thank you for coming So Yeah, so I think what we've tried to do is set up we have the contributor code base We have in our in our community. We have a community. We have a pretty big Community of I think about 190 folks that are contributing to the project and that's very driven a lot by red hat product management And myself as a community manager making sure all of those folks But what we've done is in order to sort of separate those conversations not keep them separate from each other But to encourage other communications is to create this thing called commons that allows the users to talk to each other And give feedback and we have like Trello boards and things like that, but we've tried to create The virtual with mailing lists for for different groups different sigs and things like that give them the cup the capability of Communicating directly with each other as opposed to being the the bottleneck that That makes sense Yeah, and I would I would like to see that sort of Initiative brought to open stack as well because I think we are we do a very good job of having the design summit and we have a separate world with marketplace and these sessions here, but I think the To have like a use all users come together as opposed to a committee of users or a foundation of Commercial vendors to really be able to have end users and operators talking to each other in on a regular basis And and give them the facilities for that because open stacks done great stuff in terms of Advancing how open source projects are run I think there's an opportunity here to to take get open stack to Advance the way the users communicate and connect as well to take this model on and move it forward All right you then I'll see you all hopefully at the red hat party tonight 7 o'clock La Faust All right. Thanks