 Let's start 10 years ago literally 10 years ago was For many people we started asked the question how the hell did Google do Google Maps? Like it if you remember the web back then like there was we all had web mail Where you would write this long mail and you press send and it would crash and your mail would be gone and And Google fix that. I mean they did so much else with good with Gmail But autosave was magical And and then Google Maps like there's no way they were loading all that map in the browser that to start they were doing something incredible That was given a name called Ajax About the same time there was this new web framework that came out built on it on what was me a new language In fact, I didn't ever see the language at the time to me. It was all this thing called rails There was this Danish guy who had a little squeaky voice and he was very excited about this thing he'd built called Rubin rails and And to me the web came alive, and I was so excited But I had this problem and the problem was even though I could make my app be awesome on my Windows laptop I Had no idea what happened next I I was cheap and I didn't really want to pay for hosting, but even if you did want to pay for hosting, you know The no one knew how to run rails apps in you know 2005 2006 and the ones that said they did would charge a lot of money relative to my budget of zero and You know I was just making these little trivial things So I had this this this life lesson that I've sort of carried on and it really does apply. It's If you want to win in production you do have to be Improduction there is a joke. I hope I saw in two to once it was the The the admin comes to the to the to the developer and says you know and the developer says well It works on my machine and the admin says great back up your email because it's going into production The whole problem. Yeah, there are so many problems. We try to solve here in cloud foundry But they're getting your app off your laptop and into production like Fractions of seconds after you've started. I think is essential Don't muck around thinking that the experience you're having a laptop is at all relevant get it into your production environment So back then there were a couple of barriers to production one was One was technical. I didn't know how to deploy rails. The other one was institutional for many people back then Rails was new it wasn't Java and They would sort of told You're not allowed to there's no we have no way to deploy this thing We don't know what it is. We don't know how that led to things like j ruby where it's like, you know what? We can hide this. Let's put it in a war file And they won't know and that's very clever I prove But for many of us is we have this institutional problems seven years ago seven and a half years ago. I was I Had that ten years ago. I was living overseas Which I guess technically I know I don't know where I live now So I was living in I discovered rails when I was in India Just to make this all a whole lot of exotic And so you can imagine, you know, I was pretty desperate for any anything that might eventually Be the next thing that got me out of my current situation Seven years ago is back in Australia. I was at the Ruby and rails meet up in Sydney which is held in a pub and It's awesome pub food. It's pretty good. And This guy called Tim Lucas Did two talks to light or two little talks one was on Sass, which I didn't care about the other was this this thing called heroku. Oh my lord. Oh He just a it was free. That was cool and B. He just did git push. It was kind of new back then a Lot of us were playing with it. We're solving some of our challenges in open source community of sharing and maintaining branches But he used this git tool to push These rails app and it was on the internet. It was like all my miracles had just become a product. That was free. Oh They were so exciting. So We started at a little web consultancy and then so we started using this for everything. Well, not for everything that would Like you don't shell Excel spreadsheets or anything But right notes to your wife or anything you use it for web apps But every web app now went straight to heroku. In fact It was so simple and so implicit that we'd share the app like that was the original git repo Before we then go and put it on what became github or anywhere else like it was easier just to leave it on Heroku is as a source So I was very excited about this whole experience and so having had the personal pain. I am a developer I like making HTML Become interesting or you know the whole experience As I've gone further down the stack in my profession that led me to this point now I have had to stop paying attention to the JavaScript frameworks I Can tell no jokes don't get me wrong. I can tell those jokes. I know why JavaScript is funny But I've had to stop paying attention But I became super fascinated and my mission really I grew this mission and so when I'm sorry That reminds me I get to see my slides in advance that's now. We're not going to make those mistakes. So It's okay. So heroku The wonderful thing about heroku in terms of winning in production was that you were in production on day one Literally rails new app get a net get add get commit push Basic rails that was running really the only thing that we don't have now and that story doesn't have that I believe should be that it should have happened through CI Like from the day you start a new project I'd rather every project start in CI and got deployed before you put your first bespoke commit in Like just that's that's my story for the future of our you know what we're trying to give to it to developers So anyway, but here we are with with heroku So work from day one Our consulting clients were super happy because they could now see if they can interact They could share with stakeholders And and financially it was free to get started over time. We discovered some longer-term barriers to being on heroku at the time It was still a new platform and as you can see from cloud foundry this stuff is complex and so they would have issues but Unlike heroku Unlike cloud foundry heroku is private You can't see the code back then they didn't have build packs. So you didn't even know how it was running And so that was a problem You then go to institutional customers and they were still not very excited about this public cloud concept and The the going from free to not free became expensive very quickly Rail Ruby on rails applications have one attribute that their hosting providers love and that is they love Ram That's if you're a hosting provider you pretty much in the business selling Ram and So so they loved hosting rails apps and Java apps not so exciting to host node apps Which don't just tend to have flat usage. So Should you wish to get into the business and wondering who your customers should be go after rails apps They have no idea how to run they like that don't care about efficiency and they'll just pay you more money Good business to be in All right, and so that led so I was talking in public Mostly about rails and things and I just enjoyed sharing things. I found it was I I would go from one side of the Planet to the other to do a talk at a conference to talk about some a third person's thing that I used that I thought was you know making the world a better place I just like sharing the the things I find and so I was invited to come to America Which which is here and I need you to understand that that For all the ways that you are mixed in and diverse and and wonderful group of people This is how the rest of world looks at you You don't have 11 aircraft carriers for no reason You just like to park them around like police officers. Just letting everyone know This is what we got And then when you come to the rich diverse group you find these sorts of characters He makes me laugh so much Um, anyway, sorry, I love America I do but if I if I had a green card if I became a citizen, I would still put that slide up. He is he is so cute So the people engineering yard who were I guess in the same business as Heroku But one of the original people that helped make rails app successful They asked me to come in and be part of their mission and I was super excited It's perhaps similar because I was living in Brisbane, Australia Which is you know as we discussed yesterday. It's just not America. It's somewhere else It doesn't matter where I Mean it does to me. Well, it does to the airline pilot as long as he knows we're all good and It's funny how they like to tell you where you're going That's a Jerry Sunfield joke. I'll leave you so I was invited to go to engineering yard And it's a lot like being if you're a banker or in finance and being invited to go to Wall Street It's like there's there's this there's this brand to it. You get to go to Silicon Valley I didn't know that San Francisco wasn't Silicon Valley. I found that out later And I found out many things later That turned out to be not the shiny aspect of like how expensive it was So to breathe the air on 24 by 7 basis here and due to real estate prices And my wonderful salary after expenses and taxes and everything But it was the two years I spent there I the contact with you know Thousand production customers as I said mostly Ruby and rails applications, but production nonetheless of web apps This is people's businesses livelihoods running a production and the work that the support organization did the platform engineering group The configuration management group the rolling out the challenges we had rolling out patches Whilst at the same time look, you know, we had many challenges and I just want to go through them one was You know, we wanted customers to have this experience of dev staging a prod environment being the same But we we found that hard the way we built the platform is that idea comes that have came along Is that's what happens once you have your product? You start to see how people use it you start to have the next idea And this is certainly a next idea as you want people to go through a pipeline But our platform made it we didn't really have a nice way to extend that idea to customers They could bring up a staging environment, but then we didn't really make it easy for them to push the configuration of what isn't staging to production We didn't really have a great dev story. I guess like a local laptop version of our product We tried different things over time keeping them up to date was was not consistent Another challenge we discovered over time and I say we because I'm just gonna sort of borrow the entire corporate history of Engine Yard here and much of us I learned from our our lessons I was really interested in the life history of Engine Yard like and and many stories like why do we do this and if you follow The path it was the you know one of the founders made this choice and and it was undoable we use Gen 2 and Which has a lot of benefits we could own the packages, right? There was exactly that it was exactly what we wanted to distribute a lot like Bosch Except we didn't have any of the tooling like we have with Bosch. We shipped one AMI in five years and Everything else was patches and you know at deployment time in front of customers We we had a leaky platform abstraction we sort of we were one of the first People groups to use chef like literally engine our cloud was being built simultaneously with Adam At chef Adam ops code building chef like Ezra and Adam were working together on what was to become chef and so you know that the idea of platform abstraction was a long way off and so teasing that out was really challenging and if we can't tease it out was really hard to provide services I Guess the prime example was a engine X configuration file That was where a lot of conflict happened we had our ideas of what should go in it the customer had their ideas of what should go in it and And they would win because they're the customer and so we was really hard to move that that Bundles a bunch of little attributes of our platform that came hard to move forward because of these points of conflict where we didn't have nice clean platform abstractions The other lesson I Got a feeling that perhaps due to cost due to whatever we wasn't really customers weren't really encouraged Implicitly to do the right thing and deploy on day one you would save that up For when you were ready because you know it's you're going to deploy a cluster of machines It was cost and and that I don't think was the best thing for them. We weren't encouraging them to do the right thing in my belief About four years ago Almost exactly four years ago on April I sort of mentioned this earlier today. I was I had some friends Yeah, there'd be a bit of Twitter Back and forward friendly banter between between competing companies and there were some people at that this company called VMware I was never a VMware customer because I came from Apple world. So but I knew they existed I knew everyone else was scared of them and I would banter with some of that with some of the people like James waters Who ultimately goes goes on in his own story to to to you know almost run pivotal and and so a lot of them had great Success and there's another guy called Dave Macquarie who was because is one of my good friends and and he Would be telling me about this thing that they were about to open source called Bosch And I and I felt like it could really solve a lot of our problems And I went into the demo and as I said and the next day I invited myself It never I didn't know that everyone else who was in the Cloud Foundry ecosystem at the time Used to beg and ask and and want to please can we come meet your developers? I just went over I never occurred to me to ask So and they were lovely they were really nice and they showed me a demo of Bosch for the first time And what they had was in one terminal one one side of the screen was was a terminal where they're running Bosch commands the other side They were showing v-center, which I'd never seen and And what it did was they'd run a command he changed the size of a disk and he replaced the stem cell in The same manifest change and he ran Bosch deploy and over on the other window It went through the sequential changes to automate that and I was like having a small party inside my clothes and it was It's like this is It's dancing without moving. Well, no, you're all thinking This was so exciting and so I had sort of ignored Cloud Foundry at this point And I still really and for the rest of my time at engineer through the end of the year sort of put it to one side, but I did try to evangelize Bosch Publicly and internally. I mean, I couldn't stop. I was very excited, you know They could ship stems. Oh, they had a whole build tool change for building base images Not just on Amazon, but on vSphere and later on open stack and everything else. We had nothing We have to beg permission to To you know, the old IT thing of trying to convince that one guy who knows how to do something to prioritize your request We had that with e-builds for the packages for Gen2 We had two guys with a sort of an online pipeline of what they were going to work on and it seemed to take forever With Bosch releases you can do 95% of making a package work I mean, you can go into production and then you can go and ask for help Like or how else could I do this? What was the flag I could have used, you know There's such a different model you can use of taking expertise and spreading it wide as opposed to putting it all through that bottleneck All right, so when it later on eventually I Was not successful. I had run out of had run out of Political capital Yeah, it was tough time and so I left and But I was so excited about Cloud Foundry and and I learned that that the VMware and was going to move this cloud Foundry thing into this other organization I met those people Rob me and James Bayer and a whole bunch of people that went on to lead Cloud Foundry and and I was so excited and But you know, this is the things I was learning about Cloud Foundry in terms of what it was going to be like to win in production One was it was all open source if you needed to know what had just happened. It was all there This was exciting Institution running in your own data center. We're still three four five years ago The idea of running your own data center was a foregone conclusion. Of course, you were what was the point of being a CAO if you couldn't go and build a data center There's a what else are you supposed to be doing with your job? that's It's like corruption in India's like that's the points like I want that job so I could be as corrupt as my boss And I don't mean to just say India. There are a lot of corrupt countries Like America, I mean, it's not called corruption here. It's called lobbying, but that's cool So just trying to There's a whole chain of these and I'm just trying to stop and not not say them out loud Stormtroopers, thanks. Let's build a wall to stop them. I actually read many years ago Donald Trump's one of his first books and He was he was really successful. He was doing really well. He said and There's a good book so You know, not only was it open source, but there were vendors and not just one but many coming along and And the next part was super important As a developer deploying the experience was going to be the same for five years I mean you gotta think I used to joke that a rails app because that's what I used to do rails New was how you started a new rails app and it was so simple to get started I had this feeling that it should come with a with a health warning Like in yellow blinking text Warning you are creating a production app. Do you wish to proceed like just because it's easy doesn't make it right? But it is easy and it's our responsibility to catch them all and and make and help make you know Our developers successful to let them Create success in their organizations for their end users and ultimately, you know I strongly believe that the world CEOs in the next 15 years are going to be heavily populated with people who came through our profession We have to let that happen For their own company's sake so a lot of winning in production that came from Cloud Foundry so I left engineer it started stuck and Wayne and It's a pure metric that we're doing well that you know This is more representative of everything like this conference has grown the number of conferences grown the number of people that want to participate We are a magnet for you know a path to success So one interesting thing was just in you know this I don't have access to everyone's deployments of Cloud Foundry But just looking across some of our own customers. They're like 5,000 apps deployed by the end of last year Like within that year many of them started earlier on the year And so there are just so many little metrics of success that I can have my opinions all day long so Like we didn't like in the front So there's a man called Wayne Seguin who his job at engineer in the day used to be helping to deploy rails apps Like a customer would come along to the engineer.com website and say I'd like help deploying and there was no magical script It was handed off to Wayne Wayne wrote the way magical scripts like back in the day. It was all manual and I didn't touch any of these I mean not one person in the ops team from Sarko Wayne or G or Swiss come over that we don't do this They do it themselves. This is an incredible future. It creates incredible problems, and they are different talks When you've made deploying apps easy You now need to make every other attribute easy as well and self self-service All right, so then story goes on I moved to Australia Somewhere in the last five years. It started being called Strayer, which I think is pretty cute That's that's not a picture of my house That is as representative of how bad the internet is So I miss America and your your internet And I started working on sort of production things that you know would sort of Perhaps the ideas we've learned over I mean this story ten years seven years wherever you want to start my story We've learned a lot of things and we're trying to find different ways to help when we're a consulting company You only sort of get to help five or ten customers and all their users at a time and to my mind and our mission That's not sufficient. So we're going to try different ways to try to help more people That's pretty cool. So we have like a few minutes about I don't know how long we have I've no alive We're going to do a quick demo for the people who have not seen a CF to players want to go through and talk it through This is a Rails app Because many of you Java people and I'd like to expose you to diversity So this just happens to be where I'm deploying it this is I guess Pivotal's ops manager And it's you know what it's free to download and use and play with so if you want to play with something. It's kind of fun All right, so we we're going to build a rails app. I Cannot all right, so It's this is this is an experience you want to make so simple for everyone else that you're not there I Mean I don't know if I can say it at any other different number ways This is what we're trying to do is to let people move on with their jobs without creating tickets asking for help And and to you know, you just count the success by looking in CF curl And if you've ever used CF curl, it's a wonderful way to get the data from inside Cloud. Andrew Thanks, Jamie All right So we're making a little rails app and this is a screencast so I can't skip this bit that you don't care about But I did this screencast for the benefit of a group of Ruby people and they love seeing Ruby and HTML fly across the screen It's their little thing And I wish to take it away from them Puma is is a framework that's pretty popular in the Ruby space now for being efficient at hosting your app So it works locally that's great But as I said, we need to get off our laptop as quickly as possible Otherwise, you're going to start thinking that you're massively powerful laptop is somehow representative of production Which is we're not giving you that much capacity in production So This is it. You just point it at your cloud foundry where whichever one it is and it's done It is just incredible and if it's not done Let's say you have a problem Now you know about it. Do you remember how painful it is when you package up in an envelope your code? And okay, that might be you know, but you hand it over the wall to the IT group They don't know what to do with it. They give it back if that is ridiculous There's a whole line of empathy that I espouse give developers More empathy it's a room for more empathy for the end user for the operations experience. They're the best people to fix so many problems So here we are Is the logs that you've all seen if this is new you happy to sort of show you other examples afterwards So many people seen this before Well, we'll go through a couple things because I want to tell a story so as we go through We pause here for no reason Now this is I mean as as as the guys the spring of nature guys are saying No, no need to set up DNS. No, it's just works SSL TLS is just done not my deployment, but you know you as operators you would turn on HTTPS please You know the the the build packs we can write our build packs in such a way to teach our users what to do next Right, so the Ruby build pack teaches users to say go and install this gem Why well currently no one can see your logs because you're writing them to files and That's one of the things all our users discover first is they can't see their logs because they're writing to files We need you to write them to standard out standard error. There's a simple way in Ruby We just tell them what to do so we try to educate them to solve their own problems It's asking us to be explicit about what we're running. So let's do that and this you know These little hooks of how we can communicate with our anonymous users And we don't you don't know when they're doing this right you can't catch them Like when someone decides they're gonna start a new web service, you're not there to sort of go No, no, no, don't do that, right? So we have to make this as as resilient as possible And and let them be a little bit successful and let them be more successful over time, right? Our warnings are going away And if and if your build pack that you your team doesn't that users like whether it's spring or whatever doesn't have lots of nice help at it I mean, that's a wonderful place to participate in cloud foundry is to flesh out the build packs and make them Make your users more successful. I love the CF scale command Conversely, I don't like it when the old IT departments would say, all right, you're about to deploy an app Let's just all relax How much capacity do you want? Like what are you talking about? I don't know if anyone's gonna use this yet I have such low self-esteem that I am not ready to answer your question So you start small now it is a good idea to start with multiple because then as they get killed and restarted Do all those have one running? I still don't think today. There is a way to set a default of two So that's something to teach your users about why to have to logs. Oh my lord giving you developers back their logs Can only breed empathy for those who are going to read them in production Logs tell a story of what just happened. All right after you want to know about it Because you don't look at the logs. It's like visiting a doctor's office You're there because something's wrong The only time you're looking at logs is something's wrong. You need them to tell you a story at that time I would very much like developers and PMs and and and the stories that go you in your backlog to have a And this is what the logs should look like And over time that that level of empathy Well, I think will grow into teams as that dev ops loop Which is a dev ops is a communication and team aspect more than a fancy word Builds out these ideas of what what do we need from developers so we can do a better job? Or Developers do you want to be on call all the time? How about that? No no great them do some decent logs? So they can see their own logs Doesn't make it any easier by the way if your cloud foundry doesn't have a nice place to collect logs I like to use paper trail. It's just a sass thing. You don't need to ask permission I mean you might like to ask permission, but you can just go and create an account and quietly Connect it into your app at least you can see your logs you can be you know you can win The paper trail does have a nice set of cloud foundry documentation on what to do using the Set CF cups that stands for create user provided service So that's the mechanism which you can register You know sys log end point so all the logs will drain out and Hopefully over time we start to see service brokers Streaming service instance logs through that as well so you can tell that full end-to-end story We bind it and cloud founder. It's just a special mechanism for telling cloud foundry to stream logs for your app and you can collect them Because you know at the game once at the time that you want to see your logs They're not going to be available, you know, so you want to store them You may have your own favorite a sass or you may not think of using sass Hursted sys log, but it's ridiculous to wait for someone set up elk Like you you've got a production app. So this is a workaround until you have something you like I mean also, I like this experience the cabanas Experiences and I'm not a big fan There's our there's our app Again just just as a developers think through the story think through the user story of what support going to be looking like All right, the last thing I want to show in this demo is is the SSH command and we'll skip that and get the last slide You have CF SSH it means you can get inside one of your containers and Have access to the environment, which might mean you can run one-off tasks like my great databases and those sorts of things So if your cloud foundry hasn't upgraded to Diego CF SSH command is one reason that's the best reason, you know The actually value-added reason to move to Diego is CF SSH All right So in summary, we have winning production. We want to elevate developers Don't you've got so only so many great people stop doing undead for ancient heavy lifting. I borrowed these from from Adrian from from Netflix His are the ones on the right my other ones on the left Turbo charger developers speed and in all our marketplace is what wins fastest empathy Let you encourage developers to run what they wrote and be responsible for it and not enough people have talked about Docker yet As the buzz died down. No one's I don't think any of the key notes. I say the Docker word You can push Docker app. So if you need to prove Cloud Foundry is awesome Show the demo and then go back to never using it again, but so All right. Thank you very much. Have a great day You