 Thank you, Berry I just have a question one of the things that Berry mentioned is is that HP is hiring and You know we're hiring the foundation I Wanted to know how many of you out there? How many of your companies are hiring for open-stack positions right now? So we're doing our part for the unemployment situation, I think All right, so moving on we have our our next keynote is is brought to us by Rackspace and Troy toman is a senior director there with their The open-cloud engineering team Troy is is also an open-stack board member And I think he was going to to mention he's been involved in in a very wide variety of ways and open-stack over the Over the time that it's been around so excited to have Troy So Troy toman Well, good morning. I feel really honored to have a chance to sit up in front of this room and And really speak to a community that's done some pretty amazing things over the last two years I think we're showing that there really is a different way to do this and Proving that the power of the community is is truly, you know greater than the sum of its parts I Feel a personal attachment to this this whole story of how it's come together because Of the roles that that Jonathan sort of mentioned Two years ago when open-stack was launched. I was toiling away on database automation software inside of Rackspace But felt compelled about the message of open-stack. I had been at Sun and watched what Linux did to Solaris I had ran the search business and inked to me and watched what Lucene in the open-source space did for search infrastructure in terms of making it available and immediately You know thought we'd found the right path and this was a compelling direction to go So I kind of warmed my way into the job of running the ozone team at Rackspace You may you guys may know that that that group that's been contributing to to so many of the projects that are here back in January 2010 with the mission to sort of pulling it together and figuring out how to take the second largest public cloud and convert it over to open-stack and Since that time I've gone from sort of being a newbie to open-stack to Being an open-stack contributor an open-stack user an open-stack operator an open-stack board member And I guess now I get to add open-stack keynote presenter to that credentials So I think it's a testament to how this community can work if you want to get involved and you believe in the cause you can do amazing things in a short period of time and I just think you guys for giving me the opportunity to be part of that And I want to talk a little bit about sort of that journey and and where we are at Rackspace So Little history lesson will start off with kind of how Rackspace got into open-stack and some context Particularly from the engineering point of view Which may be new to some some folks Then I want to sort of talks a little bit more in detail about how we're using open-stack We've talked a lot about products We've launched but I want to kind of dive into a little bit of of how we've done that and how we use and run open-stack And then finish that up with The chance to kind of talk about what I think some of our challenges are in open-stack in the future And really a request of the community about how to think about some priorities for next year So for us open-stack started from an inflection point in our business by 2009, you know, we had built our cloud files offering with inside software We had put together other pieces of the cloud through acquisitions And really had had had great success I mean we are we sort of inarguably we had the second largest public cloud sort of starting from scratch But at the same time we were kind of facing a wall. We really felt like the software we had Didn't have the legs to get us where we wanted to go the long in the long road And so we had a decision to make about where we wanted to go whether we wanted to continue to try to take the stuff that we had and kind of cobble it together or if we wanted to start with a blank sheet of paper and We made the decision and you know, I think a credit to the leadership at rack space They really gave the engineering team the opportunity to start with a blank sheet of paper and tell us that if we wanted to build The cloud right what would you do? and and we really wanted to gravitate towards the some key concepts which is Open source being the number one piece That we really wanted to build it in the open We wanted to do it in a community environment and we wanted to do it as a transparent of an environment as possible And then we also wanted to do some things around how we use cloud technologies and how we actually begin to dog food the very things We tell customers to try to do And and you know, I think We we grabbed that it was sort of a bold and an ambitious idea and to be honest with you I don't believe that we actually had any idea how big this would become so fast, but it's been you know quite satisfying so You know, I don't know how many how many people in the room were in Austin for the very first summit Do we have people here? There's a few hands. I can't claim to have been there Chuck who raised his hand got a cameo. He wasn't expecting up here on the slide You know, I think there were 75 people there I was talking to Jonathan the other day and he actually mentioned that the first summit It was actually before OpenStack was officially launched They largely did it just to see if people were really interested because they weren't sure if they were completely crazy This idea of coming up with an OpenStack summit So we started two years ago in this small room in Austin, you know driven by NASA and rack space You know and then fast forward Boston You know, that's actually sure we got the dates right on that But anyway, Boston was actually in October of about a year ago again the Diablo release You know which at the time was sort of the first time I think we all began to feel like that the OpenStack platform was coming together And you could really start to touch it and use it and have some dependency on it You know, it was when racks based launched our alpha product for for cloud servers based on OpenStack And I think critically it's when racks based found you know made the commitment to Create an independent foundation and truly get OpenStack out and on its own And so here we are a year later and look how far we've come with Folsom I mean, it's hard for me to believe that in some ways we thought Diablo was rock solid compared to where we've come in a year With the capabilities and the evolution of the various components that went into Folsom I mean, we've seen the emergence of quantum as a legitimate networking platform and really the basis for the future We see sender Evolved to sort of give volumes and storage. It's it's sort of it's key place in the infrastructure And we've seen just the stability of the overall project go up increased security evolution of Keystone, I mean, it's sort of blinding how much has happened in the last year and for me it was sort of illuminating and going through this talk to put this together because Probably like many of you I spend every day thinking about all the things that are still broken that aren't working right that are moving fast enough But when you think about the progress from that room in Austin two years ago to now It's actually pretty dramatic and I hope you guys take a moment to feel how awesome that is too because it's really this room That has created this phenomenal movement So here's an interesting thing. I'm gonna do something that's kind of the opposite of what everybody else does I'm gonna brag about the fact that our numbers from rack space are shrinking So in Essex we were over 50% of the contributions to that release we were about 30% of Folsom And this has nothing to do with rack space being uncommitted or slacking This is about the community Growing up more people getting involved other companies contributing making this a real community effort And so while we're happy to have jump-started this this process I'm even happier to see that more people are jumping on board and that we're really seeing a broader distribution of input into the product Now, you know part of this also Was that we got a little distracted with this thing called a product launch and we probably haven't been as community focused as we want to But you know in the spirit of sort of following what everybody else is doing I will say we are staffing up and our commitment to To continue to get engaged in the community and to redirect a lot of our efforts to more community outward Facing things is stronger than ever But I hope that our percentage continues to go down as a sign of the overall health of the community And so this is another thing that we're pretty proud of and then the last piece just to kind of level set where we are I Know we've talked a lot about the foundation, but you know, this is something Any one of the guys on my team will tell you we have been clamoring for the independent foundation for OpenStack as loud as anybody and You know for us to actually be in the you know to for me to have actually witnessed us signing that you know Rackspace signing the agreement to license all the rack space Pieces over to the independent foundation To actually be on the board and be part of the process And look, you know trying to figure out how to make an effective 24 member board is a process I think all of us are working through and we know we've got some work to do, but it is incredibly exciting To sort of be at this chapter and to see that this has a long history that's outside of rack space And so, you know, I just think our commitment to open source is stronger than ever and I also think it's been validated this community is stronger than rack space could have been by itself it's stronger than any one of us could have been by ourselves and Just looking at the growth of the number of people in this room and the number of hands that were raised the other day that we're here for the first time I think we've got a very bright future ahead of us, and it's extremely compelling So I'm gonna shift gears for a minute and give you some insight about what we've been building Or actually I guess what we've been deploying that you guys built and and sort of how how we're doing things So at the at the summit in San Francisco We announced sort of our preview of several open stack based technologies And I'm happy to say that we followed through on on all those commitments. So we are now running We're offering at least four products based on open stack from rack space We sort of launched our flagship cloud servers based on open stack on August 1st Obviously, we've been running cloud files, which was the the basis for Swift and continue to invest and involve that product We've launched our cloud databases product Which actually has Nova with open VZ is an underpinning and the red dwarf Sort of container stuff that goes on top of that as a clear basis there and we've released the rack space private cloud download a codenamed Alamo To make it available for people who want to download and deploy their own private clouds So I want to dive into a couple of these Just to share some insights maybe some learnings and a little bit of our approach Because I don't know that we've done a really good job of talking about how we actually are users and operators of rack space We have a tendency to spend time in the design summit talking about our contributions But let's dive into the cloud servers piece a little bit So for cloud servers that the sort of basic building block in our starting point Is a pretty common Probably deployment model with a couple of nuances that I'll talk through but we use Nova Obviously as the compute manager We actually I think we're one of the first groups to probably deploy quantum in production We may be the only company that actually deployed melange into production And I guess I should have added incubator Ptl as a temporary title for me for a while But you know, we obviously saw that the right the right long-term answer is to merge IP management into quantum And so we'll be following that path But right now we're still sort of running both quantum and melange in production We use glance for image management backed by our cloud files Swift based offering There and so we start there Nova obviously has the ability to manage, you know several hundred hypervisors Maybe even up into the into the thousands But one of the challenges you have when you run something at our scale is We really had to be targeting tens of thousands of nodes in a compute region to get the kind of cloud scalability that we wanted And so some of you have probably seen we're at the talk that Chris Barron's gave the other day But we started gosh, I think Sandy Walsh who's sitting there started with an initial concept called zones a very long time ago We didn't get that quite right. We evolved into cells and we actually Have deployed the cells code that we've talked about in our production deployment And it and it will go into I believe it will go into trunk and grizzly We didn't want to direct disrupt the rest of you and fall some with that that code But we actually deploy multiple cells In all of our deployments I think we have at least three cells in every deployment region that we're in today with it with our public cloud offering and so That really kind of forms the basis of what we call a region we take all these products together We then deploy Currently we have three regions in the rack space public clouds our services exist in an ORD region DFW and London And so that gives us sort of a worldwide footprint in terms of where you can go and how you interact with these So they appear as a single endpoint with sort of multiple Nova cells and these other services that kind of sit in behind them So that's a that's a real quick overview and there's not a lot of I guess magic in that I mean it really is a testament to sort of the the quality of the software that's there that we can do this But you know I mentioned in the beginning open source was part of our strategy We also were committed to trying to do some other things in a different way in terms of how we approach Approach this and one of the things was we realized that while we were building the cloud And while we were you know selling the cloud and telling the rest of the world that everything should be deployed on the cloud Our first-generation cloud infrastructure was all deployed on physical hardware In terms of the control systems that ran the cloud And in fact our earliest Deployment for I think the alpha was in a similar fashion as we were building these things out But we realized what we really wanted to do was leverage the cloud in our ability to operate the cloud And so we we call it cloud on cloud At rack space and it really is kind of our own version of inception So our public cloud servers offering the the nova API nodes the database the rabbit servers All of the control systems that you know of that run those systems run on a private Cloud built out of all the same systems So our API nodes are actually open-stack instances running on a private open-stack piece when you're hitting those And this was kind of a shift for us it was it was difficult I mean some of our ops guys will tell you they weren't big fans of the idea because they knew how to do HA in Physical gear, but it turns out HA didn't work quite the same way in in the cloud And we had to act to go out and reach out to members of the community like the guys that has tech So to help us figure out how to do this, but I'm happy to say I think we feel like we made the right decision Because it's done two things. It's given us a much more flexible infrastructure But it's also taught us a lot about how the cloud needs to get better And what and learning how to operate in the cloud and I do think you know much like we talk about dog fooding This is one case where it's actually benefited us greatly In terms of the kinds of patches and changes We've been able to submit back to the community and some of the knowledge will be able to share with with the rest of the Organization we've actually taken this concept to something we call I nova which is an internal nova offering That is the basis now for a lot of the rack space services that we're delivering beyond just the cloud infrastructure So we're starting to use this for all of our own QA environments We'll be using it increasingly for production environments But we really want this to evolve to the point where rack space is running on open stack Not just offering open stack and so that's been a pretty exciting journey for us and one that we're very pleased with So that's why I get to claim both user and operator because we are using it in both contexts Another thing that we chose to do That usually causes eyebrows to get raised when I walk around and explain what we're up to Is we chose to use a continuous delivery model for how we're building and deploying our cloud So we don't use Essex or Folsom or milestone releases. We pull from trunk Usually at least once a day and run through a pipeline that validates that We have a QA environment. We have a pre-production staging environment We have a series of automated tests that run through that system and so we can actually have a new trunk based Deployment of our public cloud software ready in under an hour in our staging environment to go to production if we want to Now that's when it works. I will tell you we don't have one every hour and Seemingly nobody really wants us to deploy every hour But we've been deploying pretty regularly about every week or two Since we launched the product and I can actually tell you that right now if you're using rack space cloud servers You're running off of trunk as of October 10th So we're up. We've already got grizzly Now you might ask why we would do something crazy like this we had a lot of discussions early on and You know it really came down to two things either you work off of a stable base and you upgrade Infrequently or you sort of do this thing to track very frequent changes up at the very beginning And we opted for so that sort of this fail fast fix fast mentality Partly because you know if you look at what's going on in the cloud today This is what some of the most advanced groups are using and I'll be honest with you when you're running tens of thousands of Hosts and VMs in an environment. You cannot test what's going on in production So we have decided push to production frequently push small changes be able to fix them fast And that's the the mode that we've decided to operate in There are still days where we wonder if we're not crazy, but overall I think It's actually benefited us in terms of staying because we had to make those changes anyway And it's a tough choice. I think vish told me the other day. There were 60 database migrations between Essex and Folsom So you get a choice you can take 60 at once or you can take 60 in small chunks but you still got to take 60 and This is an approach that's actually been working well for us and I'll talk more about where we might go with this down the road So then the question is okay, so really how well is this working? You made some marketing announcements You put your press release up sure I can log on you know where are you guys and So a couple interesting stats I wanted to pull and just talk about success overall So we realize on on August 1st One of the problems is we had to flip a product on that was going to have 45,000 customers day one Which is a little nerve-racking if you haven't done that before and so, you know We did a lot of testing going into this we realized that by August 1st between our alpha beta various internal test environments and systems that we had built over a million servers on Nova And so we actually had a lot of confidence that this thing was working because we had pounded on this to that degree And you know that actually paid off well in terms of the service So we launched this on October 1st and since then we have taken a hundred and twenty million hits against the cloud servers API In two months, so this is getting used We have thousands of customers that have built thousands of servers Across the board almost every single metric that we track The open-stack based platform is as good or better than our first-gen platform that's been in production for years Which to me is a good product launch The one I'll give you that that goes with that as we've done a hundred and twenty eight hundred twenty million queries with ninety nine point nine Seven percent availability now for us availability is anything other than a five hundred error It's not just up or down time and this includes scheduled maintenance So all those every two week deployments that we're doing had this kind of impact on our API Which far exceeds quite honestly my expectations, but I think is a huge tribute to what we're building here So you can decide whether you think this is ready or not But I think the numbers speak for the cells at least for how we're doing this at rock space So we've even gotten to the point now where we're starting to actively convert our customers over So one of the things that's nice about how this whole system is set up is you a select number of our customers And this will be increasing over the next couple weeks You can go into our control panel and choose to create an image of your server And there's going to be a new option that actually allows you to create a next-gen cloud server image So you're going to be able to take a first-generation rack space cloud server take a snapshot And then turn right around and build that server in the next-gen cloud And we're going to start actively promoting this because we have absolutely no reservation that this is the way for our customers to go And in fact our customers are voting with their feet if you watched our growth So our growth as we launched this product is actually probably ticked up a little bit from where it was before But our first-generation product is basically flatlined All of our new customers all of our growth all the things going into our system are going to the next-gen platform And so we couldn't be happier as an open-stack operator or user right now with how this is working for us and how this has come about So the other area that we've been spending time on is in the private cloud We have a lot of customers that do come to us and say well I want to run open-stack and obviously there's lots of distribution choices That are out there, but we came up with our own distribution called Alamo that you can go to the website and download Our real focus was to try to get something that was as vanilla open-stack as possible Just so that customers could easily get to get it used to what sort of there and in trunk So we've not put a lot of stuff around the outside of it, but we've tried to make it simple to install Easy, you know really the way of sort of giving you the benefit of some of the learnings that we've had in the public cloud about How this stuff gets put together and what's there? But it is really about enabling the reach of our cloud offerings to expand beyond just what's available for people that are in the public cloud Because we do know a lot of people actually spoke to one of our public cloud customers They have Essex deployed for their development environment and they feel very comfortable knowing that they're doing development on Essex And then they're pumping that stuff up to the public cloud to run their production workloads on and so there's a lot of customers That want that ability to do some experimentation locally and sort of interaction And so Alamo was a was an important part of that that process for us So that's out there as well, and it's actually also You know done everything that we wanted it to do so if we kind of go back, you know We have our three public cloud regions since we've launched the Alamo download We've actually had downloads from 125 different countries and every continent Yes, every continent Jonathan you can add Antarctica to your list We had four downloads of open of Alamo in Antarctica So we really are taken over the world 25% of the fortune 500 is downloaded it We've had over a hundred universities and research organizations that have gotten into it It just goes to show that that the open the interest in open stack is real Now the other interesting thing is There is a certain public cloud provider Who's not involved in open stack that gets a lot of discussion? They have a conference coming up where they're featuring a bunch of their reference users I can tell you that 10% of their reference customers speaking at their custom at their conference have downloaded this as well So somebody else is thinking they might want to open up their options Now we're not done yet, so obviously we announced the products that were there, but we've got some new stuff coming and The first one of these that we've we've talked a little bit about it's actually in what we call preview mode So we have customers running it in production today We'll be opening it up more broadly in the next few weeks, but we'll be our block storage offering So we're going to be launching block storage. It is cinder based So we will be we're using the cinder based implementation right out of the bat We'll have two levels of storage Available sort of a basic level and an SSD high performance level Our early preview stuff is going extremely well We're starting to see some performance numbers and scale numbers out of this that are very exciting So for those of you've been working on the cinder project. Thank you We look forward to having this running at production scale very very soon The other pieces that we're going to that we are in preview mode with that will be making more broadly available is our cloud networks And so I mentioned already that we use quantum today We use quantum for our base infrastructure when you spin up an instance You get a public network connection and a connection to the rack space service net Our cloud networks offering will actually allow you to create your own private networks and begin to link servers together across You know think of it as a private cloud-based VLAN And so our work with quantum and and with NYSERA on the back end to do software-defined networking will take its next evolutionary step again, our preview is going extremely well and You know in the next few weeks you're going to see this broaden and its availability And we're really excited about what that does and how that sets the stage For some of the hybrid computing that we've all been talking about so stay tuned for that and Then I would say the other area that rack space is really investing in is trying to share the knowledge And our experience with this in a way that makes adoption of open stack easier for everybody So beyond the download of Alamo this week we announced two rack space cloud SDKs These are not just for the rack space cloud. They're meant to be open stack driven and based So we've been contributing work to the J clouds Initiative and have sort of packaged that up and do a very accessible Java SDK We've built some PHP for PHP developers to get access to the cloud We have more of these SDKs on tap We are and we are trying to stick to the open source Mantra here of really looking for existing open source projects and contributing and making sure they're staying current with the latest developments in our Cloud and with the rest of open stack I think Everett Taves Who's probably somewhere here in the room is going to be talking more about leveraging these Particularly with the Java stuff that he's doing in a session this afternoon. So I'd invite you to go see Everett's talk on that The other thing is we're investing a lot in training We constantly get people asking us how do you do this help me understand help me learn help me do things And so we we are continuing to offer more and more training for open stack capabilities We're trying to take the knowledge that we've gathered from our private cloud Engagements from running our public cloud and make that as accessible to everybody Because I we honestly do believe that open stacks about all boats float if we can make open stack adopted everywhere Everybody who's participating in this ecosystem will benefit and this is yet just another way So we're you know, we're targeting it towards operators We're targeting it towards architects towards DevOps personnel and really tailoring some of the knowledge that we have and trying to share that as Widely and as broadly as we can in the marketplace So I just want to close but by talking a little bit about sort of the future and maybe some of the challenges that we have in front of us You know rack space actually didn't get into open stack just to rebuild our cloud It led us to open source, but it isn't what we wanted out of open stack I'll be honest with you it would have been much easier to build our cloud where they went You guys sitting in the back of the room doing it our way These design summits have a lot of debate. There's a lot of time in review processes There's a lot of overhead quite honestly operating in the community But the benefit of that is to make it better But also to make it bigger than we could have ever made it by ourselves and we are absolutely Committed to the vision of open stack And in fact is I was putting this presentation together one of our marketing guys was going through my script And he was like when you say we do you mean open we open stack we the board we rack space What do you mean? And I'm like exactly Really if you ask anybody in my engineering team, we see them all as one and we really are committed to making this happen So as I said before we've probably been a little bit distracted from the community because launching that product and supporting 120 million queries takes a bit of focus, but you're really going to see us turn our attention back to the community You know we're you know my engine This is as much for my engineers as anybody you guys are going to get more time to get involved in the community more Time to get out in front of the things that are going on and we're going to be Both staffing as well as just reorienting our priorities to get more involved with that But I also think there's there's some things that we need to think about as a community about who we want to be We started off with open stack being the ubiquitous cloud operating system And people want to overuse this idea of Linux of the cloud But I think there's a lesson there that I want to come back to but you know We did the easy part the operating system of the cloud needs to have compute management storage and networking But you know Chris Kemp brought up a slide the other day that said it also needs a whole bunch of other stuff to go around it And I think we have to think about how we move from this base stuff of the obvious piece to all the other things that could be in the cloud There's a middle ground I think somewhere between it's just servers block and networking and Any open-source project in the cloud ought to be part of open stack and we haven't defined that very well I think we can learn some lessons from Linux in this regard And I think we need to have us the idea of a core set of projects that are stable and very conservative You know the question comes up is if you've got something new that you're thinking about core It's like would you put that in the Linux kernel? Is it mature enough? Is it wildly adopted enough? Now the challenge in that is we don't really have an alternative for things in the middle that are innovative I mean, I'm looking at what's going on right now with salameter stack text and apps the metrics and monitoring stuff I mean that is absolutely essential that we figure that out But I also don't think open stacks in a position to pick a winner that says oh, we're gonna do this project And it's only this project and it's Certified so how do we think about this ecosystem where we've got a core that is solid conservative We know is the right thing but are still enabling an innovation under the open stack umbrella is something that people are inspired to Be part of now. I don't have an answer for that, but I've raised this at the board We all know working with the technical committee. We need to tackle it But I think as a community we need to raise our voices about how we want to do this Because I think otherwise this thing will get away from us in a way that may not be to everybody's benefit I think it all goes back to this notion of stability. Obviously we feel like the releases we have are stable But we do need to think about what stability means along on the road And and I think of it the best analogy I could come up with is it's like a car We want to go fast. We want to go flying down this curvy road, but we don't want to do it in a you go so You know, there's a reason these cars are engineered these way There's a reason there's this attention to detail. We need to figure out where this balance is I Personally think 60 migrations in six months might be a little much. I think we need to figure out how to stabilize I think we need to think about how data migrations are done in a non disruptive way It takes a little bit more to add a column now and subtract a column in a later migration But man from an operator perspective That's a lot better place to be than having to sit there and wait for an hour of downtime while data shifts from one Column to another I mean there are things like that that we all know how to do that We just need to slow down and take the time to think about and I don't know if now is exactly the right time to lock it down I'm pretty sure it is for a year from now, and I think we need to start moving that direction And so I just raised that as sort of a caution and an area for us to kind of focus on Now we're gonna do some things from rack space to try to help with this as I've mentioned We've got this pipeline. We want to push more of our tests up the stack We want to get them higher and higher into the integration pipeline so that they're seen closer to review Closer to review time and not you know when rack space pulls from trunk and we realize there's a problem So we're gonna work on that we want to expand the reach of our own testing It's nice that we've got this pipeline, but as most people will tell you it is narrowly focused around the rack space public cloud use case So we're not really seeing if things are breaking the KVM hypervisor because we use Zen For instance, or even our own Alamo private cloud stuff So we're gonna be broadening our pipeline so that when we're going through these gates We're starting to think more broadly than our narrow use case And the last piece is we want to figure out a way to provide some visibility. I Don't want to gate trunk based on our checkpoints But I think the community would benefit from understanding when stuff downstream at rack space is broken because of something that comes through And we and we have to figure out how to do that But that I'm absolutely committed to us opening that up and providing more visibility back to the community About the things we're seeing so that the community can begin to embrace this and it's not just sort of our challenge So those are a few things we can do inside of rack space, but I hope as a community will rally around others The last thing is I think it's time that we really shift our attention to delivering on this idea at the open cloud a lot of this year And and rightfully so has been about proving that an open-stack deployment will work I mean this is obviously what we've been doing trying to get these clouds out there trying to get them deployed Everybody here is doing the same thing right? We're having open-stack clouds pop up all over the place And that's an awesome thing But what I constantly hear is gee the benefit of this is everybody's running open-stack. It should all work together Well, I think we've got a lot of gaps there and this is the year to start figuring out how we're going to close them Look rack space is one of the worst offenders of this So I will take my shots for that Alamo and the public cloud don't work together any better than anybody else's open-stack cloud works with anybody else's So we are absolutely committing resources to get unified behind a code base and a direction and pieces Inside the rack space realm that will make our private cloud software and public cloud software interoperate But I don't want to stop there. We need to solve. This is an open-stack problem We need to solve this as a community again I think this is on the board of the technical committee as the leadership of the foundation to figure it out And I don't have all the answers yet, but it is something that we're absolutely committed I would absolutely ask all of you to rank as a high priority for this year There's three things we're going to spend time on that we will obviously work across the community on as well But continue to focus on API compatibility Obviously the SDKs help You know, we have base APIs that we're all there, but I think we need to begin to think about How we talk about compatibility of rack space clouds needs to move You know, we need to think about how extensions move into the API that everybody needs to support I mean, there's there's a whole process beyond this baseline that we need to work on We're gonna do that for our offerings. We'd like to do that with everybody else's Another big area for us to focus on is interchangeable images You know the good news is that our customer can run Essex and then push to the public cloud The bad news is he can't just take the image that works on Essex and make it work in our public cloud Now that's because we don't support a glance API. So we've got to work on that. We also do things differently We use Zen server. We use Zen store to auto configure. We don't use sort of whatever, you know, some people use metadata service We are we are going to converge on a single mechanism for instance creation and an interchange format between Glances that will absolutely ensure that any glance can export to any other glance and then any base image can boot in anybody's cloud That supports it now. We're leaning towards config drive as opposed to metadata service But that's really a community discussion that we'll engage in because we want to go where the community wants to go But we're gonna make that investment in our public cloud to make this so that anybody's images on open stack can easily get into our cloud The last place I think is just the evolution of true cloud networking So getting quantum out is a start, but there's a lot more to do there You know, we you know, even if you get quantum on both ends of the pipe There are things around Federation. So if I have a shared network L2 segment between two quantum's and two public clouds How do you decide that you allocate from IP address subnets and not conflict? I mean, these are low-level geeky things that we have to work through But they're the difference between making this easy for people and not And so these are three areas that we are actively committing resources internally to solve and and hopefully that's something That we can get the rest of you guys to rally around All right, so that's it You know, thanks for your attention. This has been an amazing journey I think for all of us and for me personally, there's there's a lot to this I think a big thing that we all have to learn how to do is take the leap of faith and trust in the community We probably all have those moments where we're a little bit afraid that if we if we put it out there If we don't hold on to it if we don't protect it a little bit We're gonna lose control of it. I can tell you after two years of doing this quite the opposite has happened The only places that I've seen rack space make mistakes Some of these were mine and some of these were others, but where was we didn't trust the community? I've learned to do that my development team does that we're gonna do more and more of that if you're here for the first time Trust this community. We're all in it for the same reason. We may disagree on the means sometimes But I think we're all motivated by the same ends And if we have two years that are even close to the two years we've had before this We're gonna have you know one heck of a thing to celebrate a couple years down the road So I look forward to being there with all with all of you for that and anyway, thanks and have a great morning