 You let me know. OK, cool. We're almost ready to go. We want to make sure that you get an opportunity to have a print copy of the user survey if you would like it. So would you raise your hand and wave at my colleague Edgar or Sheila down there if we're missing you? Oh, I see a gentleman over in the far corner. We have a few more up here if you'd like them later. And because it is 2 o'clock and we believe in starting right on time, we will get started here. Ready to roll? Awesome. Friends happily working? All right, I'm looking for a thumbs up from our sound guy. I'm going to have you guys stand on that side of me, please. Yes, we have a thumbs up. It's wonderful to have those bright lights on your face. So I will say hello. Welcome to We Love Data, 10 Things the User Survey Reveals About OpenStack. And to start out, I'll introduce myself. I'm Heidi Joy Trethewey. I'm with the OpenStack Foundation. I tried really, really hard to get them to put Empress in my title. It didn't work out, so I'm senior marketing manager. But I get to be Empress of the User Survey for a couple of months a year, which means spending a lot of time with the wonderful user committee team, designing, developing, distributing the survey, and then analyzing the data, building, reporting, and even design. So soup to nets. This has been a fabulous opportunity to work with Edgar, Sheila, and then also John Prule, who's also on the user survey and wasn't able to join us here today. So now I'm going to turn it over to Edgar and then Sheila to introduce yourselves. Thank you. Hello, everybody. Thank you for coming here. My name is Edgar Magana. I work for World Day. I'm the Cloud Operations Architect. I'm part of the user committee as well, engaging the community since 2011 in Santa Clara Summit. So I feel like all time are here. So it's glad to be around. It's glad to present to you all the reports and all the analysis that is behind this survey. So I am really happy to see you all here. And my name is Sheila Sabi. I work for Comcast as the community development lead, been working on OpenStack since 2012. And I'm really excited to be here to share the survey results with everyone. Great. Well, as you're getting seated, if you want to grab or we will bring you a copy of the user survey here at the end of the presentation. OK, so we get to tell you these 10 things. And we're going to leave a lot of room for your questions as well. So let's rip through the 10 things that the user survey reveals about OpenStack. For us, the first thing that was really surprising was this idea of similarities between the last survey and this one. So you may already know that OpenStack has been doing a user survey since April 2013. This is the seventh user survey. And we compared the results of this survey to last survey, which was six months ago right around Tokyo. What we found was that the results, like the specifics of cloud, of deployment decisions were surprisingly consistent from survey group to survey group. And we also saw that there was about a 36% crossover between people who answered the Tokyo survey who also answered this survey prior to Austin. And if you look just at those who registered deployments, the crossover is about 39%. So a surprisingly small amount of crossover, but it means that we have our arms around what a community of almost 2,500 people just in the last six months are saying about OpenStack. So that's really helpful, powerful stuff. So Sheila. All right, large, small, east and west. The users that conducted the survey spanned across 76 different countries, and 538 cities. So it's very big. 24% across Europe and Asia. And it was equal for both regions. And business of all sizes are adopting OpenStack. So from small to mid to large. So this is something that has happened in the community. As we start using emerging technology, we start using different hats. I will put myself in that position because I'm a start as a developer. Now I'm in operations. I'm doing community engagement. So in this graph, you will see we have a very new, shiny role, which is the cloud architect. It's the most prevalent that is showing right now. In addition to that one, the CIO IT is also new in these links. Something that I really want to highlight in this graph is we have 446 application developers showing there. So this is not just for operation. This is not just for infra people. OpenStack is making an impact directly on the application developers. And they are showing their interest based on the results of the survey. It was 169 increase from the last year. And that's an amazing number that we are so glad to show to you. So we also asked about project adoption rates. And we saw across the board higher project adoption rates. Let me show you what that means. And for those of you who are already squinting on this chart, it's also in our print report. But in general, what we saw was that last survey, six months ago, the top six projects ranged in adoption between about 78% and 85%. And now that number is between 83% and 97%. In fact, the top five projects are over 90% adoption. So that is with us taking into account both those clouds that are in production or full operational use, as well as those that are in the QA testing stage. What this chart excludes is those that are in proof of concept stage. But we think that this is a very exciting chart that demonstrates the maturity, not only of OpenStack as a whole, but of so many projects within OpenStack. Well, my favorite topic in technology, this is I really wanted to talk about these containers. It's a hot topic, everybody wants to see what's going to be the connection, the interoperability. I read the other day in tweeting, take a shot every time that you interrupt interoperability. So I'm going to get drunk by the end of the day. So as you can see, containers is the most prevalent thing over there. Everybody wanted to get engaged. We know all from the technology point of view it's going to perform very nicely. It's not going to affect the current performance of your application. So now we need to figure it out how actually we're going to put it into the ecosystem. So if you may ask to yourself or to us, actually how this information is valued for the roadmap for OpenStack, well, we transferred all this information to PTLs, to the technical committee, to actually take it into account for the road maps. Another thing that is a very hot topic and we are not surprised to have it in there, in second place, is SDN NFP. We just don't want to connect between machines to network. We want to actually have it connected to network return functions in a very easy way than obviously using the same set of APIs that we are coming right now. For at the end, I want to highlight that actually, bare metal is new in the least, but it's not a surprise, because now what we're going to connect is containers on bare metal. A lot of people use containers on virtual machines. Now we're going to move to the next level, though actually running directly on bare metal. And this is why people wanted to use things like Ironic to automatically deploy first bare metal and maybe put something else on the top of that. So, Edgar, a little bit more on containers. We were asking folks, not only if you're running a project in production or on testing, but also if you were simply interested in running a project. And we got responses across the board and this proves out containers again. Exactly, so we will go back to, okay, now which product would you be interested? And that reflects, it correlates perfectly what kind of information we're getting from our users. So, Magnum is the project that actually is getting more attention because they want to see. And you can see the other services, like for instance, Designit, the people who is very familiar with running up open stack in production, you may know that DNS is kind of like a big problem right now, so Designit is trying to solve that problem to provide the API to actually communicate legacy, legacy infrastructure with your virtualized infrastructure and that's the reason why it's in that position. And then we have the other projects like Ironic as mentioned before, bare metal is very important. So we see a very constant maturity in these projects. I will expect for the next six months when we present these results, you will see that maybe this Magnum and Designit will start making a little bit more in the graph that Heidi presented before in the maturity and adoption in not just POCs, actually production and QA environments. Yeah, so users are releasing new releases faster. This time some of the releases, yeah, so there's the chart. As you can see, some of the releases get higher adoption rates than others. Ice House is an example, you could see it's larger there. So this chart shows that Ice House was more popular than Havana, for example. In the last survey, Kilo, Juno and Ice House were the top ones and in this current survey it's Liberty and Mitaka, or I'm sorry, Liberty and Kilo. Actually, something that we noted was it was those three most recent releases in the last survey that pretty much everybody was running and then it's really compressed to just two now, which I thought was very interesting too. Definitely. All right, now let's talk about users adopting more than core. One of the new points of data for us or rather new ways of analyzing this round was on the idea of the average open stack cloud and how many projects they're running. So I asked our data scientists to dig into this and just provide us a count. How many projects does a typical cloud run? Well, the average cloud runs 11, but as you can see from this distribution it's more of a nine is more typical of a cloud. And then over there on the right-hand side of the graph 49 different clouds are running more than 15 projects. That's kind of like taking that long tail and squishing it together, standing it on its head. But that's also really impressive to me. It shows a lot of clouds are being very experimental and reaching beyond core services to adopt quite a lot more projects. This is a stat that you heard in the keynote, which was that 97% of users indicated that standardizing on the same platform and APIs that power a global network of public and private clouds was a top five reason to adopt open stack. And for context, because you think, oh, okay, well, you know, a lot of them are rating it maybe not their number one choice, but to give you context here, in the last survey cycle, only 60% of the users we surveyed indicated that having standard APIs was important to them. So we have a massive number, greater of people who say this is important to them. And then as you can see in some of the other business drivers, either something is an issue for them, cost, for example, either cost is a driver or it doesn't even make the list. It's not important to them. So we're excited about this idea that open stack APIs are becoming the standard for enterprise infrastructure as a service. If you let me add something, I think for people who was like probably three years ago in one of the summits, I remember one of the hot topics was like, are we chasing easy to APIs or should we create our own? I mean, it's clearly demonstrated that we are creating our own roadmap. We're creating our own APIs. We're not following anyone. We want compatibility. We have orchestration projects that actually will help to orchestrate the same kind of APIs deployment between different kinds of clouds. He is one example of that. But we are driving these as our own destiny and we believe that we're gonna put more emphasis into the technical committee to actually help on the externalization of the APIs even more than what we have for right now. All right, let's talk about user satisfaction. How many people are familiar with net promoter scores? Oh, this is good, about half the room. So, Edgar, typically if I said, oh, hey, I took a test and I only got like 50, you'd think to yourself, I get an F, right? But actually in NPS, a score ranges from negative 100 to positive 100. So, a middle score would be about zero and typically you see middling scores around zero. Anybody wanna take a stab at what the BBC's NPS score is? British Broadcasting Service, media? Just guess, what? 13 was a guess? Anybody else wanna guess? So, their NPS is five. And the average for the software industry, that NPS is 19. So, I had just said some context for the number because I have been seeing in some media reports which I'll remain nameless, that when reporting the user survey, they're mentioning the NPS score of OpenStack and saying, oh yeah, OpenStack, it only had a NPS score of 41. Well, that's awesome. It's really exciting, actually. And I think what's even more exciting is looking at how the NPS score is changing over time. And so, that's what I really wanted you to take a look at here. Excuse me. So, among people who have actually taken the time to file deployments in our user survey and you can see the number of responses here up top, they are increasingly having greater satisfaction with OpenStack. The NPS climbing nine points from a 32 to a 41. And then if we take a step back and look not only at those with cloud deployments, but all users who answered the survey, so roughly 1600 people who took this last survey, we also saw the NPS score go up by nine points. And again, it's that wonderfully consistent increase. Last cycle I have to confess, we actually reported that the NPS score had gone down last cycle. And it turned out we were comparing apples to oranges. We were comparing those deployments to the entire community's scores, which was wrong. So, we went back and we actually went back to the absolute basic data and had a single data scientist use the exact same methodology to be able to really compare apples to apples on this. And we were really excited to see the results and to see that consistent increase. So, that's our net promoter score. All right. OpenStack deployments are maturing. Okay, so. Sorry, there we go. That's okay. So, as you can see, more clouds are officially in full blown production now. Almost two thirds of clouds are actually in production. And that's a growth of 33% in just one year. So, I think it's really cool that it's moving so fast. Yeah, and that was kind of the headline for us with our press release and a lot of the communications around the user survey, that increase in deployments and production just since Vancouver. Excellent. I mean, what else can we ask, like part of the community to have a platform that actually works, that actually has been tested, proven, it's out there. There are so many different kind of companies. You're so telcos, you're so enterprise, you're so SaaS. A little bit of one of them. So, it's actually out there. And we came here. It's very amazing how it's growing. It's growing because we help each other. We help a lot. I remember changing information with other people about the networking, the storage, and all these struggling points. If you are interested or you're struggling in actually deploying or understanding or just getting your head of what to start, either the ops or any of the mainly leads or any just talking around with the colleagues around the summit, the conference, you will find the help that you're looking for because what we want is just to keep increasing these numbers and that graph to keep going to the production as much as possible. Yep, all right. So, we wanted to remind you that you're able to get a soft copy by downloading it on the website. And something that's new this year is we did a little four minute digest where we took this presentation, literally the shortest amount of notes possible from this presentation, recorded it into a video that you can find on OpenStacks YouTube channel. So, the 60 page digest in four minutes. We also are providing space for you to have comments and suggestions and perhaps even shape new questions or different questions that might appear on the next user survey. And that would be at the user survey work group. And then also, would you talk a little bit about the user committee meeting? Sure, so we have a user committee meeting on Wednesday at 1.50 PM. We'd love to have you come. We are looking for input within the user committee and the user, I guess, yeah, the work group. And thank you all for doing the actual survey. We really, really appreciate that. We wanted to show you a few credits of all of the different people who contributed to the user survey, but we also promised to show you a preview of a tool that allows you to be the data scientist. And so I will show you that next. But Edgar, did you want to speak a little bit about the user's committee? Yes, absolutely. So first of all, a big thank to Heidi and her team. She did all these possible, I mean, all the people that actually in this slide, it's contributing some weight to this user survey. Everybody and around the community, just by submitting their responses and give us all that feedback that we were looking for. But obviously it's a lot of work to actually, I guess, all that information to run on a little bit of analytics to understand. There are some technicalities that actually we need to back up for to make sure that we were making clear. I'm pretty sure that we can improve, that we can enhance this survey, that we can extend it for the new projects, for the new interoperability. Second shot that we're gonna have around this week. So please join us on Wednesday about the user committee. We have a lot of working groups. So the user committee is the way for the foundation to actually welcome the operators, the user, to hear their feedback, to hear, to understand where they want to drive OpenStack for the future. We have multiple working groups to help us to scale all these huge goals, all these different pillars that we can build. We have, for instance, the product working group. We have the user experience working group. We have the user survey grouping group. So there are so many places you want to know, how can you, from the user operators point of view, want to be part of this community? Wants to engage, help, participate? Join us to the user committee. Join us to actually providing your input, providing your health, and we will do the best to actually welcome you warm to our work. Great. Thank you. All right, just a little bit of show and tell while you're thinking of your questions, because we definitely want to save time for questions. But this is the beta version of the user survey report. Credit goes to a couple of foundation staff, Jimmy MacArthur and Wes Wilson, for designing this, building this, and right now, as a beta, I want you to kind of take it with 10 grains of salt. Keep in mind that we have all of those people listed on the last slide who are working on that print report for you, validating data, saying, hey, if you say that you have exabytes of data, that's like all of the print works, or all of the words written in every language throughout history, that might not be the right number. And like having to go back to individual users and validate data. So as you're looking into something like this beta, keep in mind that it hasn't gone through a committee review process. This is an opportunity for you to play data scientist and to actually use those global filters that you're seeing at the top to explore how the user survey might reveal other things to you that are questions we haven't asked. So here are some more examples of how you might use demographic data or organizational size data. In cases where we might have very few results, we're very concerned about protecting the confidentiality of the companies and users. And so at a certain minimum number, typically around 10, but in some cases it's higher, we generally would not return any results. We'd simply say you need to remove a filter so that you have a larger sample size to again, protect the anonymity of the user. Here's another example. We have multiple sections for you to explore just as we do in the user survey itself. And then also you can explore with cloud size, where for example, you can apply a different filter to looking at cloud size. And then finally, I wanted to point out we have loaded in data from both the last survey, the one October 2015 and this survey, so that you can switch between the two or do some data comparisons. And then I just gave you an example of one of the dropdown boxes on the global filters. There's actually about six different filters. But that URL will take you again to the beta and once more I will give you the disclaimer. Beta, beta, beta. And if I have to say beta more times, maybe I should do the shot. We can actually extend it. So we would be thrilled to answer your questions, whatever they may be, and we have saved you 15 minutes to really set the agenda for us. Questions, please. Come on, great people. Okay. Would you please step up to the microphone so that everybody can hear you and also so that there's a microphone just here. Please. I can repeat your question, so you wanna repeat? Sure. He asked if there are any, is there any data in this survey that give insights to roadblocks to adoption for users adopting OpenStack? Oh yeah. Particularly I would point you to the comments section. We asked people three substantive questions. One of the questions is what do you like best about OpenStack besides it being free and open, right? That's a very positive response. We also asked them a very negative question. What areas need further enhancement? That's where you're gonna get the meat of your roadblocks and I think we have about four pages in the user survey of verbatim comments and those comments aren't just, oh hey, I just happened to pick that many responses because we're talking about 650-ish individual responses per question but instead we were selecting the responses that were most representative of the response themes. So if we saw a lot of people talking about consistency then we were highlighting consistency as an issue and we were also selecting some representative comments that spoke to the idea of consistency. So I would point you there, I don't think that we would be able to provide you a specific data-driven answer to say most users say they can't adopt because X. I would also point you to the app developer section because they go deep on that as well. So we have somebody here. And would you please step up to the microphone to ask your question next? Please go ahead. Hi. I work on docs and so we've had some questions about our install guide and stuff recently which we'll be having more conversations about us during the design summit. So we noticed that you've got a section in here about what tools are used to deploy OpenStack which is great but we were also wondering if there's any data on packages versus source installs. We do, we do have a section about the which maybe, maybe the question could be misinterpreted by some people. So because we, if I remember properly when I fill out this survey was which distro did you use? And my distro would basically mean not a specific vendor distribution but what we're trying to ask is like you're using are you like the packages from Red Hat, the packages from SUSE or things like that. So I do remember there is a section about that specific. We ask if you use your own packages if you modify the packages or you use the distro one. So it's page 36. Thank you. Does that answer your question? Hi. Yes, a little bit. Okay. Hi. Talal, great effort, great report. Two points. The first point is there is no section or chapter measure the security or the information security and analyze it. So I recommend to include it in next survey. The second things or the second points, I highly recommend to include the enterprise architect and security officer in your role to be investigate or get their feedback and in the survey. Perfect, thank you. That's a great, great, great feedback for us. Thank you. Hi, Ashish Naderkarani, IDC. At a panel this morning, there was a question asked on whether the foundation has any prescriptive guidelines on which modules get mixed and matched with different projects and how customizations happen during deployments. I wonder if the survey has any questions on how users are customizing open stack deployments and what if any repercussions they face down the line when they do upgrades. You want to take it, but I remember something. I mean, I don't think we have a specific question like you gave your options to answer, but I do remember there are sections where you can actually add the comments about what is the specific you need on your cloud and that's a little bit harder to collect because it's reading and it's understanding the comments from people, but that's when we are trying to emphasize those kind of things. One question, one way of asking that question is which modules are the ones that get pulled out or put back in? So for example, if you're working off of Liberty, what modules are you picking up from Kilo and what's the customization that happens and why so? Right. Yeah, so as of right now, I believe we would just have to parse through the comments. But I think that's a good topic to explore some more too. Definitely. Thank you. Do we have any more questions? Well, thank you for joining us. For those of you who came in late, we do have a few more hard copies for you and we'll be up here to talk to you a little bit more. Thank you. Thank you all for coming.