 Hi, thanks for coming. My name is Randy Baez. I was the CEO of Cloud Scaling which was consumed by the Borg, I mean EMC, last October. Yeah, that's where I'm drinking this Diet Coke. Can you recite the 12 steps Chris? No. Oh, you found a beer. I did. Okay, great. So this is mad stacks, right? Beyond Thunderdome and we have three distinguished guests starting with Jesse Proudman, Boris Rensky and Chris McGowan and Caroline who's gonna moderate. We were hoping to get some audience participation and interaction. If you have a question you must go to the mic or your question is irrelevant no matter how good it is. Caroline how are we doing this scoring thing? How's this gonna work? So for each question they're gonna get scored. Yes. Okay. Yes. And then you'll make the decision. Determination is our judge. For the video I've been asked that I have to repeat this. So the scoring is going to be based on questions answered and I will decide whether they get a score from one to five and then we will totally up at the end and then Randy can beat them on the head based on how many times. Last man standing baby. Last man standing. You get to do good out with me in the boxing ring. I've been practicing. I haven't. Yeah I don't think you'd last long. Can we take you over on that? Back there close as well because we're gonna be probably loud. So fair warning I'm punchy. I was up till 2.30 last night working on my presentation for tomorrow which of course you're all attending right. Just nod your head even if you're not. Okay good. Alright so let's start off with the question that I wanted to ask which is can we extract all vendor lock-in from Open Stack. Is it possible to have zero vendor in lock-in Open Stack. McGown. No. Friendsky. No. Oh my god this is very not interesting. Jesse. Absolutely. Absolutely yes. Absolutely. Prove it. Look I mean I think the reality is what is Open Stack and if you look at the work that was done this week with the interoperability efforts and the certification of Ref Stack at a core basis you have the ability to say here the common APIs and you can get consistency in those API calls and so to some extent when I make a call to a Blue Box Stack or an Aptira Stack or an HP Helium Stack I can expect the same results and if I can expect the same results for certain sets of API calls then I have avoided lock-in to a large extent. Now the counter argument to that that Randy would say is that there is no such thing as full open source that you need proprietary engines whether it's storage, networking or other components to make the stack function and yes I think that's more important than that. Quiet, quiet, quiet, quiet. And the reality is yeah that's true to an extent but those aren't part of the Ref Stack pieces so yes you could have Open Stack without lock-in. Oh they turned my mic down. You could have turned it back up. I'm not happy with it. So what you're saying is that the only thing that is Open Stack is what's represented by Def Core. No, what Randy said is Randy did not say can you have everything that is Open Stack all vendor-free lock-in. He said can you have Open Stack that is without lock-in and the answer is yes it's just a limited subset of what is Open Stack. So I call this session Open Stack porn and I have a question for you. I'm asking the questions. It's kind of related to your question right so. Okay well then we have to score the first one you sure you don't respond? I can back off but. Jesse gets five points. What about these guys? Oh they to get zero? Yeah. Oh boo you suck. Okay Boris now that you're the moderator I guess we're playing. Do you want to sit down? Moderator hot seat. Go ahead tell us your question. So I have to just asking I guess for an honest answer. Jesse jokes aside do you seriously believe that because we now have ref stack that does high level kind of API compatibility testing between different Open Stack flavors you'll actually have real customers that will trust that the blue box and rack space and whoever else is running Open Stack will actually run compatible stacks and they will actually move workloads seamlessly between these different service provider stack options or you just kind of like you know. So were you there at the key? I'll let you finish but quiet again quiet quiet quiet quiet can we there at the keynote on Monday? I'm gonna let you finish but Beyonce had the best video of all time. Okay so we all saw this before. I'll tell you more. I'll tell you more. I can talk about it Monday. Moreover, Miranda's guys wrote the ref stack code. Okay so on the key is on Monday we watched digital film tree put a workload on blue box and on HP Healy and public cloud it was the same workload talking to both stacks they were able to move data across them continuously that's not a like it's a production workload like that that is something that will move into something that they actually do in the real world. So absolutely I think that is a reality and we're proud to be part of that announcement. Only if they're limit themselves to the features that are exposed in def core and in rest act. I saw this already in Zen and KVM. We went through this process with Nova and it failed. How many people have actually looked at the list of capabilities in def core like oh we have Rob okay yeah okay Rob Rob looks at his hand up characterize the size of the number of capabilities actually described in the current def core not our intentions but like what is tested today what is it 10% of the total number of 10% of open stack is compatible from provider to provider he didn't say all or none he said is there a compatible open stack and the answer is yes he just got a look at what's actually compatible but but your answer is essentially that Microsoft Word and freaking you know Google Docs are compatible and yes I move documents from Microsoft Word to Google Docs all day long. I think I just got clocked. It's not the total percentage that matters it's whether or not those that 10% is the critical percent. That was not how Rob was saying that it's not it's not the amount it's what you do with it. That was how I was expecting you to finish that sentence. I'm of the opinion is that for this kind of notion to be realized anything short of 100% compatibility is not going to work. Okay here's my flip question to that. Have you actually tried to deploy a workload to a morantis a mox environment that's open stack express and a blue box cloud environment and seen where it doesn't work. Wait a second we got to score the previous question. I'm not sure that everybody got to I mean this is not a debate between you and Boris okay. So I want to score Boris for actually asking that question and that one is a five for asking the question. So sorry bad van a white. So here you go five okay. So five for Boris for asking that question and then actually I want to give you four Andy as the moderator for dealing with that kerfuffle very very well. I'm getting scored now. I didn't even have to hit anybody so and let's let's take this somewhere else right. So it sounds like the thing we're arguing about a little bit is interoperability right and you know one of the things I said in the presentation a few summits go I don't know they all blend together now was that the reason people care about interoperability is application portability. Like you know if you actually boil it down with use cases I want to run my app in a couple different places and I don't want any surprises right. So I don't think I want to get in the weeds too much but are we seeing other than Jesse's foe use case are we seeing any real places where people have complex applications and they're able to just sequence it up on any open stack cloud or even any public cloud that's not open stack willy-nilly and not encounter problem. Isn't that called cloud foundry? That's what I was gonna say we absolutely have customers that are doing that with cloud foundry. The red head guys would say open shift exists too. You're wrong but I'll throw the name out. I didn't do that I didn't do any of that and no culpability. Go ahead Boris. Yeah so I this I mean I keep going back to the debate that we had but we do have customers. Which you lost. Yes I lost sorry. Oh you should be sorry for baiting me in there and attacking me out of corner anyway. We do have customers that actually use kind of a multi cloud strategy hybrid cloud strategy but the only time we see this working is when there is some sort of a mediation proxy layer that is in the house that is actually the endpoint for their workloads to talk to. So that can be in some situations we're talking about higher level of abstractions cloud foundry or it can be something like scalar or it can be the right scale. Yeah but but I don't even know the notion the notion of heading to millions of right now. I mean it just seems like a general problem. I don't even see people doing on Google and Amazon. The problem with that though is that once you start going down that path you're not using OpenStack anymore. You're using your users are using the abstraction layer. They're using the OpenStack core services that have hit the ref test cases and are certified. OpenStack is about providing primitives compute networking and storage and now we've added all these additional capabilities on top that are trying to compete with all these external services but at the reality if we just pay attention to those primitives those primitives work well. They work well for private deployments. They work well from premises deployments. They work well for hosted private cloud. They work well for public. Jesse your average application developer couldn't give a shit less about which is why you need a platform as a service. They're deploying their application. Everything else is just magic. Yeah. Jesse I think you need to get some allies on this panel because it's like everybody against you. No that's not how I roll. It's like it's Matt's why it's beyond ThunderDome. All right. Christopher gets a five. Again. For that statement about applications or what the developers are actually doing. They don't give a beep about what's underneath. Shit. Okay. Sorry. I want to go back. So the reason I don't have any allies on this panel is because everybody is approached OpenStack the wrong way and they're selling it as a piece of software and it's not a piece of software. It should be distributed as a service and we've talked about this last year and I talked about it all the time but that's the reality. Like it's so complicated as a set of software that trying to ship it as a piece of software into organizations who are barely keeping up with VMware. It's a recipe for disaster and so our whole model has been shipped as a service delivered as a service whether it's hosted or on-prem and I really believe that's the ultimate path forward. If you look at the OpenStack... Okay we're doing it. We're doing it. Look at OpenStack in total. It's a medic cloud. So I'm going to pass the moderator baton to you. Would you like to ask a question of our esteemed panel? Should Chef and Puppet be distributed as a service? That's what your objection to... Well Chef is distributed as a service. You can buy Chef server from Chef as a service. So yes. Okay granted. Great. You? Chef and Puppet as a service? We actually have customers that have a notion of Puppet as a service running on top of OpenStack so sure it can be. It's a viable... But should there be a specific vendor who only does that? Back to controversy. So... There's a market for it when they don't see why not. Do you want to score that? Nope. Alright. You all get zero. Okay. So tomorrow when I do my presentation I have some data from the survey that I sent around to some of you last week. You saw it and you saw it. Well you got to pay attention to my Twitter feed man. You got to get your alerts on your DM phone or whatever. So every time I... You need to know what's up. Set to buzz you when Randy tweets. Exactly. Doesn't everybody has that right? So in that survey there were a couple of... We're gonna do a sneak peek of tomorrow. There are a couple of projects that were almost universally unloved and you give me your perspective within OpenStack and you may not use the answer. All parts of OpenStack are equally awesome or equally suckage because that is not a valid answer. But give me your answer. What is the part of OpenStack that if you could wave magic wand you would change tomorrow? Jesse. Is this my prediction about what your survey will say? No, but... Well I mean I think you look look the problems are obviously in neutron and I think neutron because of its breadth although it's getting much better. Celiometers is a big challenge because of the way it was originally architected and I think you look at a lot of it the newer services that are being developed and they're just new like that's part of it. Boris. I think that the problem with and there's there's a lot of projects of their own problems neutron etc we can talk about but I think that there's kind of a systemic issue of delineating between some things in OpenStack that should by nature be kind of a community programs and those that should be vendor driven and I think... Pick a project. So triple O versus fuel. I don't think that there should be an official deployment program in OpenStack. Alright I want to say neutron but I'm not going to. The project that I want to go away is glance. It's actually a feature of Swift that should be a component of Swift. It could be like 50 lines and we should do that. Perfect. Score us? Okay next question. I can't say this any other way. What the fuck is up with the Keystone API? Why is it not OAuth 2.0 or SAML? Jesse? No comment. Boris? I don't know the answer. I don't know what's wrong with it. Why is there a custom API? It's because the Keystone developers have never found a authentication API that they didn't want to reinvent badly. To be fair, what's their works? Like it is working. We proved you got multi... Am I wrong? It works. It's not ideal but it works. I know I can't tell you specifically what's wrong with the API but I know that the systemic problem with Keystone has been that there's projects in OpenStack that were developed by people that were real kind of deep domain experts and there's projects that were just kind of taken on and then have been the more obviously kind of evolving and had injection of opinion by all kinds of random people and Keystone is one of those cases where there's no people that are experts in actually doing authentication that seeded the project initially and as a result we get what we get now. What was it? The Grizzly Summit? Were they invented domains? Let's reinvent Active Directory on top of a restful API in Python. Okay, perfect. Let's score that one. Well, I think about my next question. Oh, there's an audience. Should we take some questions? I mean, I can keep stirring up shit. Let me let me let me score first. So, so, five for Christopher for poking that hornet's nest and then a four for Boris to adding fuel to the fire and then shall I take questions while you think of your next question? Yeah, I'd like somebody come up. Because there was a lot of controversy. I'm taking your microphone. You can't talk anymore. Yes. Any questions? Anybody? You can walk up to the mics. I can start calling on people. I know like 20% of you at least. So the scary thing is after this talk, I'm actually gonna have to go into witness protection. So if you guys don't see me again, the Keystone guys got me. Just look for a man without a bow tie. Now that I hear Keystone's a disaster, should I get an early ticket flight home? Please answer the man's question so we can score it. Are you going to fix it while you're here? That's not an answer. That's a question, Boris. No, you should stay and you should fix it while you're here. Come on. I guess not a disaster. I think we're over. We're over. We said that that we did not say the Keystone was a disaster. We said that or at least I implied that the API choice being non standard was a disaster. That's a very nuanced response to this question. And this is the Thunderdome. We don't do nuance here. We don't do nuance here. We do not do nuance here. Okay. Raise your hand if you think APIs are the same as the software. See, it's not that freaking nuanced. Nobody thinks that. They're afraid of you. That could very well be. I'm scary. Okay. More questions? All right. Well, I'll just keep asking questions. I'm prodding at this thing. So, Chris, you've got like this awesome distribution thing that you guys built that like installs OpenStack in like five minutes while Josh McKinty juggles balls. I actually have... I won't say what kind of balls. A kangaroo scrotum that I was hoping Josh would be here to juggle, but he's not. Okay. So, oh, we have a question. I already started. So, McGowan, this thing installed OpenStack, right? So, why does it now install Hadoop? It also installs Hadoop because we built an orchestration framework that we'd been calling an OpenStack distribution. So, you're not pure play OpenStack anymore like Boris? No, we're not. Were you pure play OpenStack before? That's a nuanced question and the people that Thunderdome don't like nuance. I think we were, but we weren't OpenStacky in the way that we accepted that everything was OpenStack. For me, OpenStack is Nova, ostensibly swift, glance, unfortunately, keystone. That was like a crisp specific question. I don't even know what to ask you guys. All right, he gets whatever she scores and you guys get zero. What is the weird thing in his jacket? Like, I don't even understand what just happened there for the last 30 seconds. It's something about kangaroos scrotums. I didn't quite get it. Rob, go ahead. You didn't think this was going to be PC. Do you need them? I mean, look at them pictures. So, we were... Yeah, never mind. Last night, we had a conversation about unicorns in which the horn was not a horn, but... Anyway, which sort of ties into that other conversation. I'll leave it to the reader to finish that. So, I have a couple serious ones for you all to debate. Non-Python OpenStack, alternate languages. Fuck yeah. I mean, sorry, I'm not on it. Yeah, definitely yes. Are you talking about GopinStack? GopinStack. I didn't specify language, but alternate languages. I don't know. I can't think about that. Okay, I love this question. That's not Excel file. Thank you, Rob, for asking it. Go ahead, Chris. That's not controversial. I think the answer is yes. Boris. It's definitely yes. And I think that if you look at most developers that are good, they typically work with more than one language. So, this notion, the notion of actually, you know, the only reason OpenStack has been all Python is because they say it's all kind of community of developers, and if it's all Python, you know, it's kind of we stay basically on one language. But I think that most of the people in OpenStack community are fluent in various languages, and there's no reason to actually limit to Python if the end result that you produce is better. If you code more than Python. Yeah, of course. That's like an outbrainer. Jesse. Yeah, I think it's absolutely useful to accept multiple languages. It's about the end game. What are we trying to accomplish? Right tool for the job, right? I'm not going to go right and make some Python. So what's the icebreaker that makes that happen? I think the Swift thing was the icebreaker. Hold on. Score that one, please. Trying to, we've got to have a score at the end. Three. For who? Three. For Jesse. All right. Zero for everybody else. All right. Icebreaker. And so what's the icebreaker? So I don't want to hear the the go Swift thing, because I think that's too obvious. What is going to force us? What's going to be the forcing function? I mean, if that's it, then say it, but what's the forcing function that allows that to happen? We need more Ruby in OpenStack. We did that once before. It actually is not getting in trouble. Go to C or something. There's enough Pythonese in the crowd that's just charged. I tell you, I tell you faces aren't my thing. I'm sorry. I tell you, if Google came now and said we'll contribute all of Kubernetes code and we were contributing and developing under OpenStack umbrella, the problem will be solved. But it didn't work in the live demo this morning. So was it using only the death core? I agree with Boris. I agree with Boris, though. I do think like if we could maybe invite into the project, something that could fix certain projects that we've identified that have problems that's already proven in production system that is written in Java or whatever. Java Keystone. The original implementation of Keystone. It's about currency and throughput and performance, right? And if Python is not the language to do that, look at languages that are. So could it be Go? Yes. Could it be C? Yes. Could it be C++? Yes. Could it be Java? Absolutely. Like, let's figure out what the language we need to do the performance substitute for the specific components. Okay. Scores. Rob, you got a third one? All right. So I know everyone is on the panel next year, Rob. Wait. So before Rob scores, Boris gets a five for bringing up Kubernetes. And then Jesse gets a three for saying that, yes, languages should be looked at as other options. I thought it was a three for Ruby. It was a minus ten for Ruby. Three times more Ruby in this project. If you want to ask a question, get up behind the mic. Yeah, I'm trying to help Randy out because I know I've stepped on every landmine there is. So I can certainly help target these. But go, Rob. So OpenStack will at some point fork. I'm starting you from that point. I disagree. So who will fork and why? And if you want to say no, then go ahead, Boris. I don't think that OpenStack will fork the end. I think that we're simply, I mean, open source communities are about momentum or too much momentum behind OpenStack for anybody to be able to fork it and create an alternate community. That'll be a fork. There can be kind of downstream distributions versions, but they'll never be a fork. So you're wrong because it already has forked. It's just forked internally into a number of companies, like Circus HP, where they did this giant fork and they maintain this huge set of patches and they're like, holy crap, that doesn't make any sense. And then they came back to main. If you follow that logic on every single implementation of OpenStack out there as a fork. No, ours is totally open source. Sorry, bro. So I own the domain. I own the domain. We're listening to Chris now. We're listening to Chris now. I own the domain sports.com. And the reason I'm going to fork OpenStack is so I have something to do with it. So I was talking to Siri yesterday, getting some feedback. And Terry said to me, he said, and this was just like kind of, I don't know, slightly eye opening. And maybe it's very obvious. I thought you said Siri. I did, too. I was very confused. No, watch this thing. Terry Perez, you heard of this guy, right? Yeah. Okay, just making sure you're paying attention. I was talking to Terry. Terry. I don't know. I'm not French. Anyway, so I was talking to him and he sort of described this thing that I kind of knew. I think we all know, but I didn't eat like the way he said it. I was like, Oh, yeah, that's it. He said, and the context was, should we allow different languages in or other competing projects? And he said, you know, I want people, I'm paraphrasing. Okay, so don't quote me exactly. He said, you know, people should understand effectively that OpenStack is about the community, the common values and the governance model, not necessarily about code, specific codes, specific APIs. And I, and I just was like, Yeah, like that just felt obvious. Jesse, what do you think? Do you think it's obvious that that's I mean, that's why I think we should use Ruby. Yeah. So Dev Core should be about those three things. Where's Rob? He's high. You know, I absolutely agree, which is why I don't believe it's been a fork, because there's been enormous amount of investment in building the proper governance, building the community. And there's not going to be a parallel community like OpenStack. Yeah, I think I'd be hard. I disagree. I absolutely disagree. I think OpenStack should eventually become a big tent. I think until we actually finish building something that people can run into, like download off of OpenStack.org and deploy with no Jesse, no Boris, no me. We haven't succeeded. What does that have to do with what I just said? I'm getting there. All right. If we're going to big tent everything, we need a stable base that actually works. And until we have no distributions, no service providers that are doing providing OpenStack as a service, we shouldn't be going and adding Manasca and... Thank you, Captain Obvious. Okay, score us. I disagree with him. I disagree with you. Disagree. Okay. Next person to ask a question in one second. For Boris, talking about the other distributions and how they have sort of forked. Good job. Two, for Christopher, and minus five for Ruby to Jesse. I don't think she likes Ruby for some reason. This fine gentleman here, please identify yourself and your company. Scott Copeland, Chemical Abstract Service. Operating system for the data center. I've heard this tagline associated with both OpenStack and Mesos. Is there a conflict here, or is there a right way to use them together? I apologize in advance for both of those projects for using that phrase. I am very, very sorry. And I hope they will both take that phrase and put it where the sun doesn't shine. Okay, answer. Don't blame us, blame Randy. Answer. So can I speak for 30 seconds? As long as it's 30 seconds. Okay. So I think that the notion of operating system as I see it comes from the operating system as we know it. A good example is Linux. And one of the key things that it does is that it actually ties the different components in the computer. 15 times. Into something, some cohesive system that you can interoperate with. OpenStack does the same through drivers with different diverse elements in infrastructure stack. So I believe that Mesos does not do that. So I think that OpenStack indeed can use that term. Mesos is more of a cluster manager. That's like 45 seconds. You drive me crazy. So Mesos has a really nice logo and I think they did a good job there. But outside of that, I mean, at the end of the day, there's so much more to the operating system for the data center than any of these technologies actually deal with. And we all ignore that fact. Hallelujah. Because so many of us are so many of those involved, the community actually have never run a data center that they don't actually know what's involved. Tell them, brother. Five for the question asker. Wait. Oh, for Jesse. But I do like their logo. Thank you. That's it. Okay. Identify yourself and your company, please. My name is Joey Dial. I work for Century Link for Telco in the US. Century Link? Yes. Joey. Tell us. So we have hundreds of vendors here, all offering OpenStack solutions. My question is, everyone, all of us have seen the rule of stack here that they're trying to put together. Why is Morantis the only one that successfully completed it? That's used to me. I just found this out like literally before walking in the room probably about 30 minutes ago. So you guys only want to have done it. No one else has actually completed the task successfully. What's the task? I missed it. So Intel has this thing upstairs called rule of stack and they have a bunch of computers in a rack and then people type on the computers. Also Intel is an investor in Morantis. And it's a measurement of how fast you can install OpenStack. And the reality is the Blue Box team is too busy servicing customers to daily galley with toys. In the customer's defense, everyone does want to... I mean, I'm not joking. We're opening orders from the summit. So we're not completely gelling on the Intel rack. So, Chris, you're going to use my service. Yo, if you use Morantis OpenStack inside your stack, because they won't be so busy installing. It's simply a guy in the back. You know, if I hold up my finger one, this guy gets muted. If I hold my fingers up two, this guy gets muted. Three, this guy. Okay? Thank you. Go ahead, Joey. I'm zero. Go ahead, Joey. It sounded like that was a divergence from the question, actually, but I am curious. That's like one of the major cruxes for most people trying to deploy the stack. And that's where Triple-O comes into play for. It helps everyone quite a bit. But it's not a single vendor apart from Morantis. And there have been quite a few that have tried to complete the challenge. And it shouldn't really be a problem anymore. It shouldn't be a challenge is what I'm saying. So why isn't everyone up there saying that, hey, here's my solution. Here's how fast I can do it. If you're so simple, then why not? Cut his mic, number one. That's all I'm saying. Chris, I think these two answered. Can you answer the question so we can score it? Would I be able to walk up with my installer, plug a USB key into their servers and power cycle them? Yes. All they give you is hardware. That's it. You bring your own solution, you install your own platform. That's it. Do I have to use their USB or mine? What is a CD? How do I even get my CD? Oh my gosh. They didn't let me bring my CD burner over the border. I'm going to hold you there and score it. But first, because I think this is a great question, thank you for asking it. This isn't a cookie cutter type of installation either. It actually is a somewhat segmented installation. So it's not like someone can walk up with pack stack RDO and like just hit enter all in one and you're done. I mean, you actually have to put some deployment thought into the process. Sure, but so I want to ask people, raise your hand if you think that installation is more important. Well, first we'll do installation and then we'll do operating and running the cloud. So if you think a really fast installation is important, raise your hand. Okay. If you think that like ease of use and managing and operating the cloud after it's installed is far more important. Raise your hand. Can I add another one there? A consistent installation? I think if you I think if you get this thing cloud and then you demonstrate that in five minutes and you get your forest and installs in ten hours, but then you know one of them is ten times harder to operate then I don't think that install time. I think it's more important and much more value comes out of operations, but I think that solving the installation is the first basic thing that you have to do in order to be able to then automate operation. We're cutting it off. It's 2015. Score us for that one. I'm not going to hold up the sheets anymore because they're all being chilled. So I'm going to be one as well. So I think I actually give the score to Randy for this one. It's not just about the speed, it's about the operations and yes I agree with you Mr. Audience, you can, you should be able to have both and why neither of these two are working together to make it be both? I have no idea. You can ask a question but you have to go back to the end of the line so we can get another question in. Sorry. You're welcome. Identify yourself. My name is Matthew and I'm from SAP. So this is a little bit of a different type of question. I hope you guys are there with me. So I want to know, this whole event's going to be a little bit worried. So can you tell me about the state of open source license compliance in OpenStack? Because I work in open source license compliance, like a legal compliance analyst. I know OpenStack has like a CLA and all that. But how do you know that some contributors not just like copying and pasting code from non-commercial sources, AGPL or whatever, like what types of audits or programming and tooling you use to ensure that we're compliant in contributions and such. Thank you for asking a practical question. Starting with Jesse. You can answer great. All I can think about is the Richard Stallman eating the foot off, foot thing off his foot video, GPL. Like that's all I can think about when you ask that question. I don't know. It's not my job. I think it really is a problem in the community. I think that if you compare it to many other open source communities, the CLA process is probably one of the more rigorous ways to ensure from at least legal standpoint that you're indemnified. But there is no process right now for doing audits and things like that. So it's a problem with no perfect solution. And right now, actually, in fact, there is a sentiment in the community that OpenStack's CLA process is too rigorous and you shouldn't have a CLA. You should have what is it the other one? DCO or whatever. DCO exactly, where actually there's like a page published on OpenStack.org that says well, by virtue of contributing code to OpenStack, you guarantee that it's open and it's yours. So that's if you can choose the CLA or the DCO. Yeah, and DCO is the most common way to actually contribute, which is much, much, you know, there's no, there's no requirement on the person contributing to actually perform an action. Chris. So Randy just went through an acquisition which had a compliance department that went through all of his stuff. Holy shit, you have no idea. And since there weren't any like egregious patches to OpenStack after that, I'm pretty confident that currently the state of OpenStack's compliance with other IP is fine. Look, what you in Portland like HP got on stage and said our entire value proposition is that we pay for the parties and we pay for the lawyers and we have demnify everything. So if you have worries about HP. Woohoo, HP. All right. Boris gets a three. All right, we've only got time for, for, we're gonna try again as many questions as possible. Hey, Dustin from Royal Bank of Canada. And I'm just wondering, is there an inherent conflict of interest in the way the whole project structured in that there's not lots of money to be made if it's not hard and difficult for people to download and use themselves? Oh, I love it. You, you talk too much. We're gonna start with Chris. From the audience, could I phone a friend? Yes, absolutely. Like the major contributors to OpenStack are people who have a vested interest in it being hard as held install. It was that, been that way since the very beginning when Rackspace was the major contributor, they wanted people to be using Rackspace. Great, Boris. I think that if you really believe that, then you guys should have been developing your installer in the Open like we have been doing and not holding it back. And to answer your question, I think that whenever there is, there's a problem like that and an opportunity for somebody like Piston to create their private installer, that'll make it easier. There's an equal opportunity for somebody like Norantis to develop one under the umbrella of the OpenStack community and make it five seconds. So it's a balancing kind of notion. Okay, done, Jesse. It's not hard to install. It's been easy to install for a year and a half. The hard part is operations. The hard part will continue to be the same. Same question applies to the operating. Same question applies. It's hard to operate distributed systems. Like look at Amazon, like it's not easy. Next question. We're trying to get as many in as we can before the cloud powering at the door. Anyone on this side? I'll throw you guys a mic. Giri Baswa from Trilio Data. So OpenStack is a non-profit for everybody. Semine can write things off. Hold on. Where's the question, dude? You got to finish that with a question. So it's OpenStack is not about making money. Two are false. Totally false. It's absolutely about money because like this is the story of all OpenStack software. OpenStack software is funded by organizations that can pay full-time developers to work on the software. And those organizations need to have a P&L that allows them to do that. So in some sense, yeah, you got to make money. Not in some sense. In total sense, you got to make money. How many here are developers who are also independently wealthy? I think we've only got the one. Okay. Squirt a question. I was going to say Mark Shutterworth is not here. So he would be the one that would be the developer that's independently wealthy. Alright, Rob, last question. I've got a small one to close this out if you want. Go on. So if OpenStack was a cocktail. Oh yeah. Which cocktail would it be? Come on. It's a Basil Haid in Manhattan. Stirred. Not shaken. Okay. It would be martini. It's actually just a highball glass. I want to answer that one. Because you, everybody needs it, but there's always a vehicle for something they actually want. I thought it's because you throw it against the wall. I'm going to say a Tootsie roll, sort of like equal parts orange juice and Kahlua. Try it. So I'm not going to have you do the drinks at my wedding. Who's, you're getting married? Congratulations. I'm having a baby. Thank you. I had a baby. I guess we're over. Am I over? Are we done? Well here, I need to do the results. So before his highball answer, he was a second, but with the highball answer, he is now tied with Boris for first at 26 points. I'll give you a hockey jersey for that. And Jesse scores an 11 for bad behavior. I'm going to go retreat to my golf cart in the parking garage and drive around later tonight. Nice. And then Randy for being the moderator of the Mostos gets a nine and the fellow who asked that complicated question gets the 10. Oh, he scored higher than you. He's got higher than you. Nicely done. Okay. Thank you very much for coming and I hope you had fun. Thanks everyone.