 Very happy to introduce the next speaker is Monty Taylor. He's a distinguished technologist He's a member of our board of directors and he's on the technical committee. So he's a very busy man It doesn't get a lot of sleep. So I'm glad he's willing to come up here and tell us what's on his mind So come on up Monty They doing sir Good to see you Hello, hello. Thanks for coming out. Thanks everybody. It's really exciting to be here in Hong Kong First time we've been out in Asia as I'm sure you all know and it's really exciting to see everybody's everybody's shining happy faces Hope you all are enjoying the keynote so far this morning I'm going to talk to you a little bit about open stack and hybrid cloud Unfortunately for me, it seems like everybody's already been talking about open stack and hybrid cloud So there might not be as much new information Here as I originally wanted there to be or maybe that actually sets me up into a good position Where you might actually believe the things I'm telling you so first of all Why am I talking to you as? As Mark said, I'm a distinguished technologist at Hewlett Packard I sit on the technical committee and the opensack board of directors So thank you all of you out there who voted for me. It definitely gave me more things to do with my copious free time And it's been a it's been a joy But the reason that I probably got elected to those things has more to do with what we do inside of open stack Started and I'm currently a core member on the open stack infrastructure team for those of you don't know about that That's the team that works inside of the open stack community to run all of our continuous integration and developer automation So any of you who have submitted a patch to open stack and gone through the process of getting code review and getting your patch tested That's on that's on our team You can you can totally come and join and help us out with that by the way I also the mark just gave me a fantastic intro in terms of the triple-o work which is basically work shared by a team that I manage at HP and And then the fine folks at rack space that are there with mark And and in addition to all of that as if that wasn't enough things for me to be doing HP apparently lets me talk to customers This for those of you who know me might sound like a terrifying proposition to put me in front of a customer But it turns out That it's really helpful Because it's it's really good to get feedback about what customers want It's really good to get feedback about about why they why they're interested in open stack And so I'm gonna talk to you a little bit about about what I'm hearing from them What they're telling me sort of where I think we're at on that and then possibly Some things that we as a community need to do to to address some of those concerns First of all and as you've been hearing I was gonna say all week, but it's day two So I'm not sure we can classify that as a week yet But but hybrid cloud is here. It's now it's happening. It's not just a pipe dream There's there's I've been hearing people talking about it being You know being a goal. Is it possible? Can we do it? It's it's really here It's really it's it's really here now and the customers really want it. They want it for a few reasons First of all, they they want open, right? They they're talking to us about open sack because they they want things to be open and they want they want open In large part because what they don't want is they don't want vendor lock-in I have talked to absolutely zero customers who have walked up to me and said I really want is if you guys could Build something that would lock me into a platform and keep me there Prisoner I'd really I really want that as a feature. Nobody says that absolutely zero people have ever said that sentence other than me just now But they they don't want vendor lock-in and one of the reasons they don't want vendor lock-in Or part of that is they want interoperability and interoperability is a thing that we we talk about There's also people in our community who think that it's a it's a fool's errand But it's actually really important for for some really specific technical reasons if you're gonna run Your application in the cloud because after all this cloud thing isn't all that interesting by itself It's just a way to get some compute resources, right? You want to do something with that or you want to run an application on the cloud and if you're gonna run an application on the cloud that means you've you've kind of put some of your some of your balls into play there and and what you don't want to have happen is you don't want Issues with a single vendor to potentially cause problems You if if your single vendor that you that you were using say has an outage And that's the only vendor that you're that your application is running on then it's your customers aren't going to say to themselves Oh, that's fine. It was just their cloud provider. I'm really gonna blame them What they're gonna say is wow, I can't watch TV right now And they're not really gonna like that So what you want to do as an application provider as an application Developer as somebody who's going to run your business on top of the cloud is you want to have choices You want to have the freedom to be able to design and run your application in a way that's going to serve your customers, right? You don't want to be telling somebody, you know You don't want to be tweeting that you know your service is out because there was an issue in some sort of thing You want to you want to have the freedom to actually be able to to affect The the life cycle of your application yourself and for this reason open stack is actually seems like a pretty good answer to To our to our customers to the people that we talked to they're pretty excited about this because Similar to the are going back to the the mission statement that that mark just put up on the screen It's it's our mission to produce the ubiquitous cloud computing platform for public and private clouds Right and this means that you can actually develop your application to take advantage of all of these features And we can also have multiple private cloud or multiple public cloud providers Providing a place that your application can run on So that's great like that's that's the theory that they're they're telling me that and they're like hey This is this is exciting. We're thrilled We can we can make you know applications and they'll they'll run for forever and we'll make all the money in the world And that's that's fantastic. They have a few questions. I mean They wouldn't be bringing me in if they didn't want to ask Ask me about a couple of things So these fall into a couple categories and to talk about them and then we'll come back to to sort of what I think our answers are here First of all is I might have just double hit that. No, uh-huh. It worked properly The first question is is are we really open? Is is this is this open part for real or is this just another sort of form of packaging up an elaborate vendor lock-in scheme? There's there's been plenty of of open source things that are out there that aren't that aren't really open So that's that's a concern. They've got They're worried about the maturity. I don't know if any of you noticed that we had our third birthday recently but That's not that's not the longest amount of time for a piece of of world-class software to event in existence So so they've got questions about that they want to know they want to know things about that They want to know about security Especially given various things that have been in the news recently People are concerned about about what's going to happen to their to their data's They're worried about the complexity If there are as as I basically just gonna be referring to all of Mark's slides apparently is what I'm gonna be doing As as Mark pointed out we have a lot of pieces to this to this puzzle it it turns out It's kind of complicated to run a cloud So they're a little bit worried about that. They'd like to know some answers there And and and finally they're sort of interested in this upgrade story, right? Like so you go and you install 20 data centers worth of I make gosh 20 data centers That's pretty exciting But you install 20 data centers worth of cloud and then we make a release another release in six months Are they gonna turn their cloud off and install a new 20 data? Turns with a cloud and then turn it back on that probably wouldn't be a very good user experience for anybody using those clouds so that they sort of want to know some some answers what we're gonna do what we're gonna do there and So to sort of address all of these things to To to think about answers to to those questions for them It it comes back to us it comes back up to us as as the community And so I sort of want to highlight a couple of things that I think are important there But the the key the key initial message that I'd like to get across here Is that this is about what what we need to do is the community? This isn't this isn't about what these other people who are the community need to do for me This is about what what we collectively as the community need to do together And that's one of the things that I think that we have as as again as mark So eloquently put in the in the last presentation This is This is one of the this is one of the most impressive things that we've built We've built a community of over 250 companies and you know over 1600 Developers that that contribute in over 12,000 individual members of our foundation That's that's kind of an insane number of people to be working on this thing And the fact that they collaborate all the time the fact that we collaborate all the time is is Nothing short of of a miracle really if you consider the names of a lot of the companies that are up on on the slides when we Show them And the fact that they come together every day The fact that we all fly to various different cities every six months to work on this together and sit down in rooms and look Each other in the face and make solutions to problems is kind of amazing So I think that there's there's several aspects of this of how we work together But all of them at the end of the day come back to us working together Here in this room and there in that room and well in my living room Although I don't really want to invite you all to my living room because I don't think you'd fit But but I'll be there and you can be somewhere else and now I'm babbling So I'm gonna move on to the next slide of what we need to do first of all I think we need to we need to as always we need to face the facts We need to be very very honest with ourselves about where we are And about where where we want to where we want to go next We need to be we need to be honest as I as I just said about the fact that the person sitting next to us Isn't going to solve our problems for us. It's our job We have to we have to solve our problems. We have to come together and and do that We need to we need to be honest about the things that we need we need to not try and whitewash anything over into the corner saying Oh, no, it'd be fine It's it's really important that that we we we have been honest with ourselves so far I think and we need to we need to maintain that even in the face of the resounding success that we've that we've got so far One of those facts that we need to face is that we are currently hybrid We as a project actually we actually run a rather large hybrid cloud application I mentioned earlier that we have a community run Developer infrastructure team. We've got several members on that team work for me at Hewlett Packard There's people on that team at the open sack foundation and there's other other community members And we do we do a decent amount of work because it turns out that everybody out there Does an extremely large amount of work in producing patches? So thank you. Thank you for producing all of those patches I thank you for giving us the the lovely job of actually testing all of them I think Russell Bryant tweeted a couple days ago that he was thrilled that in our community We consider something that isn't tested to be broken and I think that's I think that's right on The thing that you may not know about that depending on how much you know about how our development process works Is There's there's a couple things first if you're going to submit a patch to us It's it's gonna it's gonna get submitted up and we're gonna run all sorts of tests on it They're gonna they're gonna do everything from code style that we're actually gonna spin up a cloud Install your code make sure that it works and test it and then it's gonna go through code review And then after it's gone through code review and somebody a couple of different core team members have decided that they like your code We're gonna test it again Because things might have changed since you submitted it so we're gonna do a lot of we're gonna do a lot of testing on that We do all of that work currently in two public clouds So we actually have our entire build infrastructure spanning Two public open-stack clouds that are run by different vendors So if anybody tells you that you can't do interoperability You can't do multi-cloud applications. They are absolutely lying to you. We are doing it every day a lot You can go and look at look at Sean Diggs numbers recently. It was something like what? 20,000 clouds that we installed in a week a couple of weeks ago After the release when everybody was on vacation So we do this a lot across two different clouds It got really exciting this last cycle because of the work that the triple low guys are doing now We kind of have to figure out how to test multi-node bare metal cloud deployments And what we have at our disposal are public clouds so The first step is to figure out how in the heck you test a multi-node bare metal cloud deployment when your only resources are a public cloud that gives you either Zen or KVM instances and I Think it might eventually be possible to write something that could maybe only use those resources But it turns out that we have this thing that can deploy multi-node bare metal clouds so so we got a rack of hardware from HP and We've got the triple low team has started up a has done an installation of open stack go figure Wow, we've got this cloud thing. Why don't we use it? Um, so we now actually as a project effectively have a private cloud that we're running because we need some capabilities That don't exist for us in the public clouds that we have at our disposal So we've got that spun up and it's actually in an experimental fashion in our in our node pool application It took exactly zero changes to our development infrastructure To actually add that private cloud even though that private cloud is going to be there to test bare metal deployments for us And and so not only are we actually provably doing a multi cloud interoperable applications on a daily basis We've we've now actually expanded into the hybrid cloud world We have workloads that we run on our private and on our public cloud and we do all of this completely in the community Completely in the open and it works pretty darn well So that's number one thing to for any of the doubters or any of the haters that are out there. We're doing it. It's working It's actually working really well So so yes, you can write your applications to do that. I promise That said We We don't I don't want to I don't want us to sit on our laurels One of the the questions that customers are asking was about are we really open and I think that I think that one of the Elements of openness that we've become really good at is the open collaboration It's not just enough that the code is open source But it's it's that we have a community where you can come in and collaborate on things And it turns out actually if you sort of sit down and think about the engineering of it It's more efficient for you It's more cost-effective for your engineering team to just work with us Don't go off into the corner. Don't go off and write your own thing because you don't like our thing come fix our thing Come come be part of be come be part of the thing that we're doing and as a as a as a specific example of this I Can't even begin to express how excited I am that I get to tell the next little Tidbit of stuff and it's already been tweeted this morning. So my thunder got stolen a little bit Hopefully don't follow that Twitter feed actually hopefully you probably do follow that Twitter feed I'm gonna stop talking So a couple years ago when we started the HP public cloud horizon was Not great It wasn't really ready To be the dashboard of a public for a large public cloud So the engineering team that was there at the time said themselves. Wow, we can't possibly use this. This is terrible Let's go right our own I'm not really sure where the engineering math came across in that thinking that it would be less Less cost to go right one from scratch than to just fix them was already there But ignoring that for a second that went and so we've had a we've had a proprietary Public cloud dashboard for the last little bit Until somebody realized that that was just a giant waste of money And so what's been going on over the last cycle is we've had folks working on getting the features that we need Into horizon so that if you were to go right now if you have a public HP public cloud account And you would like to manage that using horizon you can go to horizon that HP cloud comm And you can manage all of the servers in our new beta public cloud Which I can't even begin to tell you how excited about that But not just because that that happened but because of one of the things that it shows about collaboration It's it's it's all well and good to come in and collaborate on the thing that you think is already working And it's great You're just gonna send up a couple of patches Sometimes what you need to do is you need you need to you need to think hard and about digging in on the thing That maybe isn't quite there yet. Oh, this is this is sort of a neat concept, but it's not it's not really ready yet So I'll just do something else. Well, no don't do something else fix the one We've got if you don't ours is probably gonna come by and surpass you anyway Because there's a lot of us working on this and you're probably not going to hire 1600 developers to work on your on your private thing so just work with this We're really friendly Except when I'm yelling at people but in general we're really friendly And and we get stuff done The the folks are asking about maturity I sort of tip my hat towards this one a little bit earlier but They're worried that we're that we're not mature enough potential. We're only three years old when I was three years old I Didn't do very many things that were useful to anybody I think maybe my parents liked me, but that's probably about it In this case, we're we're doing pretty well. We we we test the ever-loving mess out of this thing Like just so so much. I already mentioned 20 20,000 clouds in a week But that's that's actually that's actually a lot of testing Um as a community the the thing the thing that I'd really like to see us move forward on there Is that as much as we do all that testing and as much as that system is really great the There's really only a couple of companies that are contributing full-time Developers to working on those systems So it'd be really great if some of you out there who were CEOs or CEO CIOs or whatever Could give us some more resources on that because the team's doing fantastic We kind of keep doubling every six months and we're not doubling the infrastructure engineers So you could really use some help on that one, please Um Security is another one that's similar to that we have Security process we have security apparatus. We have a vulnerability management team We have we have security working groups. We have people doing pen testing on on open stack clouds And that all is is is working pretty well Except same sort of thing. Um, this isn't this isn't going to be a problem that we solve once and then we're done with it Uh, it turns out these people keep writing new patches Go figure. Uh, we're gonna have to keep dealing dealing with security as an ongoing issue Which again means we we might need More people in the room to come help us out come join the security vulnerability management team Um, it's it's really important, right? There's absolutely nobody out there who doesn't want this to be secure And it's it's we're we're doing a good job with this right now But but if you if you track the if you track the growth of what we're doing over time We've we've got to have we've got to have more help. We absolutely have to have more help there to to carry that into the future um and then finally Continuous deployment is is a thing And and without getting into too much trouble with theory our esteemed release manager I think that our six month releases are really important actually They give us they give us a point to to pause and reflect to figure out what we're doing a goal to look at in the future We get together here or elsewhere. We probably don't get together here every six months since the first time we've been here Um, but we we get together every every six months to do planning and that actually so far There's nothing broken with that the planning cycle is is actually pretty fantastic Um, but if you're running a large cloud, uh, as I was mentioning with the upgrade ability question earlier There's no way that you're just gonna stop and upgrade in six months That's it's it's insane to think of even at a single data center Uh, uh, even at a single data center scale much less at 20 data centers or 100 data centers Um, so so the the scale of the scale of changes has to has to be able to go in on a continual basis Uh, we've got as as was mentioned already today very excited about it. Uh, my guys working on it The guys at red hat working on it with us, uh, the the triple o project. I think is really important here uh, especially When we're looking at the fact that we've started running a continuously deployed open stack cloud in the community Like if you come and start hacking on triple o you can become a member of the admin team of that cloud and you can help run a continuously deployed open stack cloud With everybody else and and then we'll be feeding that back into the cycle that that gives us resources for the testing So it's it's actually a really nice, uh dog fooding cycle. Um That's that's not it though. Like that's that's a that's a mechanism and it's great And i'm really thrilled with the the work that everybody's doing there Um, but if you go and you look at the at the summit sessions, uh this time Because we are here to plan what we're doing in the next cycle If you look at the summit sessions, you'll see the uh the high amount of of sessions that are pointed at How do we how do we deal with uh stability? How do we deal with upgrade ability as we're looking for this because having mechanism that can do a continuous deployment Is meaningless if the code that you're continuously deploying Rakes itself when you go from when you go from from version to version So we did there were the nova team did an amazing job over this last cycle of getting a whole bunch of Upgrade ability patches and work done in we've got we've got grenade, which is our our First stab at upgrade testing. We're looking at a whole bunch of different ways to to look at at upgrade and downgrade paths between Between patches and this is a this is a thing that That I think is getting a getting a lot of resources and over this cycle We we kind of have to we have to nail this down is we have to admit that this is this is also a model We we may make releases every six months But we actually it's not okay for us to break things in the middle of those releases anymore We've we've grown up. We are actually mature enough already That it's it's just not acceptable. We just can't do it And and we're working on it and we could we could use we could use everybody's help so uh So in general to to to sum up actually and and I might I might give you a couple of minutes of your of your lives back To sum up I I want everybody to do what they can To think about us as a community not Not as the community is being they not the the community being those people over there who are going to get this done for us But the community being being us community being everybody who can show up And and participate because you you can all come on come out and participate And and to ask as you're interacting with this How can you work on the core pieces and not just your vendor plugin right because you can write an amazing vendor plugin for your for your networking stack But if the core thing doesn't actually work Then your amazing plugin is just sitting in in a big pile of mud And so it's it's it's on all of us to show up And to work on those pieces And to and to make sure that the that the entire stack itself Is is is ready is ready to go and also think about the places in the project that might not be as sexy To the product groups inside of your company things like things like the qa team things like the the infrastructure team Things like the vulnerability management team We're all we're all looking for for more help. We're all looking for collaboration. We're all we're all eager for people with new ideas To come in we're we're eager for we're eager for people who aren't me to to hire people for the infrastructure team It'd be really great And that's that's how we make this that's how we make this thing Really really saying and get to the next step So in any case, thank you very much for for coming out and listening to me this morning I look forward to seeing all of you in the design summit sessions and around around the conference Around the conference this week. Thank you very much