 Okay. Yeah, why don't we go ahead and get started. I think this is a pretty good crowd. So my name's Joe Brockmire. You guys may remember me from such previous careers as Tech Journalism or working with Novell with the open SUSE community or Citrix with the CloudStack community. I'm currently working for Red Hat. I joined them in August. I'm basically on the open source standards team, which is sort of a multidisciplinary little group of people who try to make our upstreams happy, basically, is the way I would put it. How many people watch how I met your mother? One or two? Okay, I won't even bother with that reference then, because not very many people. Anyway, when people ask me what I do, I always feel like stuff, because it varies from day to day. Yeah, it's really fun explaining it to people in the tech industry and then trying to explain it to like family or people at a party gets really interesting. So anyway, this is a talk about cloud. This is not going to give you any specifics about implementing a cloud or anything like that. This is more of an anecdotal and ideas type of talk. So I want to start off warning how many people have read or have seen season one of Game of Thrones? Season two, season three. How many people have read all the books? For anybody who did not raise their hand to the last question, there are spoilers, especially if you have not seen season three. Yeah, there's some spoilers in this presentation. What there is not in this presentation is any of the more HBO salaciousness. So some of the better Peter Dinklage quotes, I'm afraid I could not include. I will not be sharing what I am the god of or what someone may be the god of. So yeah, I've tried to make sure this is family friendly and safe for the Linux town cloud open audience. Severe. Yes, go ahead. I like honesty in an audience. I like that. If you do enough shots of good scotch right now, you'll probably be fine. I did and I won't remember this talk later, so no. So let's get started. Why Game of Thrones? So why in the world would somebody try to conflate cloud computing and Game of Thrones? Those don't exactly seem like a great combination, right? Not a lot of cloud computing going on in Westeros that I noticed, although probably a lot of cloud computing in the CGI that they're using for the series, but not any cloud computing in the actual show. I kind of got the idea because I read two series or two books. I consider a series really just one large book in a lot of ways. So I read two books and I'm going to talk about Game of Thrones, this entire presentation. I'm going to talk about the last book at the end of presentation that's not anything related to Game of Thrones. But I was reading these back and forth on planes to Australia and things like that and I sort of started tying some ideas together. Game of Thrones like a lot of fiction is not just about the swords and sorcery and the characters that are in the books. There's a lot of politics and machinations. There's a lot of historical references in Game of Thrones. And I started recognizing manager stereotypes in some of the Game of Thrones characters, especially Tyrion, or not Tyrion, but Tywin. I'd love to work for Tyrion as a manager. He would be great. But a lot of Tywin and some of the managers I've worked for, a little bit of Daenerys and some of the managers I've worked for or people I've worked with, I thought it would be interesting to talk about ideas and implementing Cloud by using Game of Thrones quotes or metaphors. I like the Red Wedding as an example of how to treat instances of the Cloud. Spoiler for folks. Yeah, the Red Wedding gets really ugly. I hope you're not fond of the Starks. So any presentation is improved by Peter Dinklage, really. My favorite part of Game of Thrones. And to be honest, I was a little desperate for a presentation idea because I'm not on the technical side of things. I don't have anything super insightful to say about deploying or implementing a Cloud that like Hugo, guys like that that actually work on things could share. So I wanted a presentation I could do. Game of Thrones wasn't the only idea I kicked around, though. Thought about the Godfather. That might have been a pretty good metaphor for everything you could learn about the Cloud, but maybe a little too old school. I thought about Futurama as I'm a big Futurama fan, but the number of times that series has gotten canceled, I thought that just might be a bad metaphor for any IT project. Got really crazy. Thought about Little House on the Prairie for a while and just tossed that idea out. So we ended up with Game of Thrones. So what I went with is some quotes and characters from the books to introduce concepts. Anybody remember this talk between Ari and Jon Snow? First lesson, stick them with a pointy end. How that relates to Cloud. There are a lot of concepts that you need to have down for Cloud. And you need to start with a POC. You need to start small. Don't think you are going to move your entire infrastructure over to Cloud immediately. Don't try to run before you can walk. One of the things that I noticed on the Cloud Stack support list when I was working a lot with Cloud Stack and in the forums last year was people who were trying to do things that had never, people who had never tried to do anything more than run individual servers or a little bit of virtualization were suddenly trying to put together infrastructure to service and they were flailing a lot because they just didn't have the background that they needed. I talked to a guy who recently left Eucalyptus and he was telling me one of the big problems that they ran into with a lot of their people was they found that because they tried to provide a clone of the AWS services they ran into a number of customers who really had written applications for AWS had never run infrastructure. That's a bad combination because they wanted to run their own AWS internally and that takes a different kind of knowledge than it takes to deploy an app on an existing AWS or infrastructure as a service. Which brings me to my second point. If you have only worked with virtualization and or operating systems, you should adopt this approach. You know nothing. You don't know nothing. You know something. But you're going to have to know a lot of things that maybe you didn't know before. You're going to have to know about complex networking that you may not know as an admin if you're an admin. You may need to know some things about storage that you didn't know before maybe didn't want to. I certainly, when I was learning my way fumbling around Cloud Stack just as a beginner, I have known my way around Linux for a long time. But I started running butting up into my areas of ignorance pretty quickly. So you will find yourself in that way pretty quickly. Which leads me to my next point. Only a man who knows, who's been burned, knows what hell is truly like. You're not really going to know what putting together the cloud is like until you do it. You can talk to other people. You can gather your requirements or do whatever. You're not really going to know what it's like until you've gotten into production. You've done your POC and you've gotten into production. By the way, I'm not trying to discourage anybody. Game of Thrones is not an uplifting series if nobody's noticed this. But I'm not trying to discourage anybody. What I am trying to do though is make people wary and understand if you are a little wary when you go into a project, you're probably going to approach it more cautiously and you're going to be ready for problems when they occur instead of saying everything's going to be just great. Because usually things are not. How many folks have been in IT five years or longer? Is everything always just great? Okay. Why do you think Game of Thrones is so popular with this audience? How many people are bronies? See? Exactly. All right. Another of my favorite characters, more so in the book than the series, although this guy is brilliant. The actor is brilliant in the show. But I actually like him as he's written in the series a little more because you get a little more into his head and you get longer speeches from him. Storms come and go, waves crash overhead and the little fish eat the big fish and he just keeps on. This is a tenuous tie but basically I'm tying this to monitoring. When you get into conversations about cloud with people who have not gotten into it too deeply, they almost always center on cloud stack versus open stack versus eucalyptus or whatever. They don't get into, okay, well what are you going to use for monitoring? What are you going to use for creating images? What are you going to use for these other pieces that should be part of a wholesome breakfast, an actual competent cloud? People don't get into those things as deeply as they should before they start getting into things. So if you're going to start doing a POC proof of concept of a cloud, you need to start thinking about all of the pieces and not just this. Another thing, I think on every slide I'm probably going to end up saying this, I'll just stop. Another of my favorite characters, Caitlyn Stark, doesn't get as much screen time, doesn't get as much time in the book but she's an interesting character. But she talks to her son about, I wish you could do whatever you want in reference to his marrying the person that he loved as opposed to marrying the person that he was supposed to, leading to a bloodbath eventually. Basically he just followed his heart. There are a lot of people who want to deploy a cloud and they want to think about doing a cloud and they kind of forget about the legacy systems. You have to remember that more than likely you're not starting out with a green field. You're coming into an environment that has legacy applications and they're not going anywhere. I had a really good conversation with a guy that works for ADP at Linux Foundations End User Summit in New York and we were talking, I'm like, well do you have any interest in cloud? Have you guys thought about cloud stack? He's like, no, do you want to get your paycheck next week? He's like, we have very complex systems that we have put together and they work and we're not going to tinker with those to deploy cloud. If you have legacy systems, you should be first thinking about do a proof of concept, do a new application in the cloud, then you can start thinking about how do I forklift this stuff into a cloud. Don't just go straight from oh, everything's fine. Let's just put our 20-year-old what used to be on a mainframe and then migrated to Linux. Let's move that into the cloud and see how that goes. Any questions or comments? Another of my favorite characters. By the way, I have a, this is not a spoiler. I've read all the books but I haven't gotten obviously to the end. My bets are on one of two people standing by the end of the series, him or Hodor. My money is on one of those two on the Iron Throne by the end of the series. So this is in reference. He tells Stark early on, I did warn you, don't trust me. This is in reference to proprietary vendors. VMware, if you're thinking about a cloud, I'm not going to get up here and tell you to pick Yucca or cloud stack or open stack. I'm going to tell you to evaluate all of them because they're all open source and you can do approve a concept and see which one works for your environment and they will always be open source and they can be forked and improved and all of that good stuff, which is why we're all here, right? Because open source is really important. So when VMware guys show up and say, hey, we'd like to sell you some cloud, say no, thank you. You guys have already shown us from your previous price hikes and license changes, you're not to be trusted. We don't want to buy into this in something that we're going to be, any cloud infrastructure you put in, it's going to be like a decade plus project, right? You're not going to put in an infrastructure to service in your environment and you're only going to be there a couple of years. And the same thing goes true mostly for applications you put into public clouds, right? You're not going to put an application that's critical or important in AWS and not need it a year from now. So you need to think about where these applications are going to live and which infrastructures you're betting on. Anybody remember this where he basically starts bargaining with the barbarians because he wants to live? This is a tenuous one, I grant you. Cloud is not about saving money. If you go home with only one or two ideas out of this talk, one of them should be cloud is not actually about saving money. It's about doing things more efficiently. It's about having more agility in writing applications. It's about having more robust applications. It is not going to save you money. Your cloud services, how many folks have gone from servers running on bare metal to virtualization? Did you see less utilization? Did you get rid of servers? Did you actually consolidate and have to run less servers? Or did you find out that after putting virtualization and you realized how easily you can spin up VMs, you started needing more hardware? You needed more equipment. It didn't save you money. It made you more efficient probably, I hope. It made things a little bit easier, but it didn't actually save you money. Same thing is going to be true of cloud. It's going to be easier to monitor the usage, but you're not going to save money. You like running your applications. Whatever the price, do what the business needs, but don't expect to save money. Now a little encouragement at this point during the talk. Do not be afraid of doing a proof of concept. Do not be afraid of trying out cloud. Make sure that you actually go into it expecting trouble, but don't be so afraid that you don't embrace it. Don't let people in your IT infrastructure who are afraid of losing their jobs to automation block it. This stuff is important and you're going to need it. Kind of echoing what I said earlier. There is a tool for every task and a task for every tool. Make sure you're thinking about things like Puppet when you're talking about cloud, thinking about the right kind of monitoring. If you do cloud write, you should be able to automate and do a lot of things a lot more efficiently. But make sure you're using the right tools. One of the things I saw again on the cloud stack mailing list is people who were running cloud stack and then they were thinking about, well, how do I have an image? How do I convert an image from VMware to Zen or to KVM? I've already created my golden image. Now how do I take that golden image, ISO or whatever, and how do I convert that to KVM? If that's how you're doing it, you're doing it wrong. You should be creating these things on the fly with Puppet or other tools so that you're not sitting there clicking or RPMing your way to all the applications and then lovingly creating an image and then figuring out how do I convert that to this other virtualization format. Make sure you're considering all the tools. Likewise, once you get there, one man on the wall was worth 10 beneath it. Once you get cloud properly implemented, one admin can be as good as 10 in an old style server room. But you have to do things right. Any questions or comments so far? Here you go. You're kind of, okay. Not to put you on the spot or anything like that. All right. One of the things about, as you watch Game of Thrones, how many people have read the books? Because this focus is on the common. The books focus on the common people a lot more than the TV series really does. Does it feel like, well, this is a wrong audience because most of you don't have to deal with this. How many people were laughing at the US government the last couple of weeks through the shutdown and everything? Are bad. Not really mine. I did not vote that way. But you get the government you deserve. So the problem isn't that the government lacks the tools to solve the problem. It's just they don't really want to because they're more focused on the divisions between them instead of actively governing the country the right way. This happens in businesses too. We see splits between different business units or IT and development or the DBA doesn't want to do anything. The security guy doesn't want to do anything because it might make his job a little bit harder or he can't verify it's as safe, blah, blah, blah. If you get trapped in the mindset of silos of developers versus the admins or any of those things, finance versus the financial people want one thing and IT wants to go another way. It gets ugly pretty quick just as it does in the books when the throne becomes up for grabs. Really it's all about each house wanting what's best for the one or two people running that house. It's not about what's best for the entire land. So keep that in mind when you're going about implementing projects in IT. There's a limited amount that we can do in IT to control that but do your part. But the king dreams the hand builds. There's another good quote that I didn't include but basically when the business needs things it's up to IT to build them. It's up to us to come forward and actually put it together. So what your managers or what your business needs is what you're going to build and that also means that you're not entirely free to build the cloud or whatever service that you feel you need if it's not what's called for for the business. Like I was saying earlier with the legacy apps if you have a couple legacy apps and that's all you need to run maybe cloud isn't what you need right now. So you need to think about the business. This gets down to budgeting and dealing with you know everything in the cloud but also has to do with one of the nice features of cloud is you can actually monitor the usage and you can actually see what everybody is using and you can actually use quotas to make sure that you're not having to deploy more services than are necessary. Another good one from the series is you know common people pray for rain, healthy children, a summer that never ends. It's no matter to them if the high lords play their game of thrones so long as they're left in peace. In other words whatever you implement needs to work okay because they don't care about you know what IT wants. They care that their services work okay. The end of the quote anybody remember the end of the quote from the series or from the books it's basically you know they don't care so long as they are left in peace but they never are. So hopefully they get a better shake out of their IT department than the people in Westeros get out of their government. Lannister always pays his debts. When you are doing these kind of things you are going to incur technical debt, you are going to have need you're going to need to have a plan to deal with the technical debt as well as for implementing the cloud in the first place. So make sure at every stage of the game that you are planning for how are you going to fix what you change along the lines okay. Changing an IT department to be more agile to go into dev ops and that sort of thing requires a lot of planning and it's going to take something away from other things for a while okay. So big spoiler alert for anybody who hasn't read the book so in the series so far we've gotten to the point where I believe she has turned the unsullied on their masters and she has conquered the city and everything and now everybody's come out and like oh you're the greatest and she's won right. It gets pretty bad in the books after that because it turns out that she doesn't actually know how to govern all of this once she's gotten to that point. Just a reminder you again need to be prepared for what the systems are going to look like and you need to make sure you know how to manage them. I just wanted to include this one. If it helps you know you guys how many folks are familiar with the pets versus cattle analogy okay well let me give you that one real quick. So if you have a system running on bare metal you have your SAP or something like that that's a pet okay I have a cat at home love my cat would spend ridiculous amounts of money if I needed to take the cat to the vet and it turned out that she needed some expensive surgery or whatever. Cattle not so much you know pretty much hoof and mouth disease boom next you know that's what your cloud instances should be like you should not be afraid to wipe out the cloud instances that's why Netflix wrote their chaos monkey stuff okay you don't want to run chaos monkey on your SAP system right that gets bad people will not like that on the other hand you know you do want your cloud applications to survive chaos monkey okay so you should name you know you should have some server instances out there when you're designing your template and it gives them domain names just put stark in there to reinforce the message. Another one here you know the night's watch is the only thing between the realm of men and what lies beyond we have barely enough resources to keep our lads armed and fed this is the IT department right you barely have enough resources and yet you're really the only thing between the business and failure right how many businesses these days can survive for a week without their IT systems any guesses I mean anybody work at a business that could survive a week without IT two days one day exactly so that is something to keep in mind as sort of a uplifting comment you know you guys are the ones between the white walkers I'm not mixing game game of thrones and walking dead am I no it's the white walkers okay I watched too many things with zombies in them so yeah you know make sure that you're prepared to be the sword in the darkness right and when you play the game of thrones you win or you die there is no middle ground once you get into implementing this either when you do it successfully or you're probably looking for another job okay keep that in mind when you get into these sorts of projects I couldn't come up with a good one for this but I thought I would include in any suggestions for hood or got nothing all right so like I said there were two series I read at the beginning of the year that really kind of influenced my thinking about cloud game of thrones a little bit less than the other one this is the other one the phoenix project how who's read this one okay a lot fewer people than I want to see this is a great book okay and I don't say that about a lot of books I it is one of the best things that I've read about it in the last five years so what this is basically Gene Kim and Kevin Bear and George Bafford wrote a novelized version of a company with an IT department in trouble and they're you know going from traditional systems you know non-automated bare metal systems to DevOps by the end of the book the nice thing about this book other than the fact that the advice in the book is good is it is a novelized version so it's not dry and boring to read like a lot I mean it's not great literature it's not John Irving or anything it's not even George or our Martin but it makes it a really easy read I read this on a flight I think from St. Louis to San Francisco okay it's a really short read it's also even something you could give to a significant other who doesn't understand what you do because it really it's readable and it gives an idea what an IT department has to put up with but it goes through the politics of having to deal with the different the marketing group and the developers and the database guy and the security guy and all of those things it's a really good approach to understanding how you should go from traditional IT to cloud okay strongly recommend this we were lucky enough to have him out to give a keynote at the cloud stack conference this year in Santa Clara and he's a great guy really personable so cannot recommend this book strongly enough and that is actually my presentation so I got through that a little faster than even necessary any questions or comments nada by the way there is a microphone here if you do have a question or comment so they can actually get it on video what's that I couldn't actually other than the pets versus cattle thing you know couldn't really come up with something good you have something I should use if I do this talk again all right any other comments or questions all right well thanks very much folks have a good rest of the week