 Okay, it's time to get started. Hi everyone, I guess it's getting towards the end of the day, so not too many people here. But that's all right, I hope you all had a great day today. Good, informative day here at the OpenStack. So how many of you are new to OpenStack itself? About three or four people. How many of you have actually deployed OpenStack? None? A couple? Okay. So relatively new people. So the topic of today I'm going to talk about is, I kind of stole this word called SuperUser. I'm sure you saw the SuperUser awards at the keynotes yesterday. SuperUser is a different kind of connotation. These are not the same SuperUser. I'll define what they are as we go through the presentation. This is really all about, more about how you make a case. So if you're starting an OpenStack project within your company, how do you make the case, the business value, that it makes sense to go do it? So I'll go through with some real concrete examples of real users and real outcomes that they have experienced. So I'll share some of that with you. So hi everyone, my name is Kamish Pamaraju. I am the vice president of product marketing at Mirantis. How many of you know Mirantis? Almost all of you. That's great. So I'm sure you know what we do. We have an OpenStack distribution. We are an OpenStack services distribution. And now our new messaging is around building OpenStack clouds, operating them from our customers. And if they're interested in transferring it back to themselves, we have an offering for that as well. So without further ado, so today this morning in the last couple of keynotes, we saw some incredible user stories, you know, companies from banking industry to large-scale scientific computing, some really, really awesome stories. But there's a broader economic backdrop that is driving all that, which people have started to call digital disruption. And some people yawn when they hear that word, but I want to give you some concrete examples of what that is and how that is impacting the market and how companies are reacting to it. So just a very quick, I'm not going to spend a whole lot of time on that, because that's what's really driving new customer expectations. And what those new customer expectations are, which we are most familiar with, I'll cover that really quickly. And if you're in this business of delivering digital services, how do you do that? What are the key driving factors? What are the key business drivers for that? And I'll speak a little bit about that. And then go into the meat of the presentation, which is how cloud leaders or how super users, as I call them, are actually delivering these business benefits and how you can do the same in your companies. So what strategies did they use and how can you use those same strategies to win with the OpenStack implementation? So that's kind of the overview of my presentation today. So I'll start with the digital reality of today. So very quickly, what do we mean by digital disruption? I guess many of us will have our own definitions. But I mean, I don't have to emphasize this, but we all live in a world that is digitizing very quickly. We are very familiar with that in our day-to-day life. Millenials, if you have kids that are teenagers, you know that they live in a digital world, right? Smart anything, anyone? Everything is smart, smart phones, smart watches, everything is smart, smart cars. And all of this is being driven by many, many trends. I mean, and I'll give you some amazing examples that you may not have heard of, people that are actually taking advantage of IoT mobile machine learning. In fact, I'm in the Silicon Valley, and as you know, Silicon Valley is ahead of everybody else. The biggest buzzword right there, right now, amongst all the VCs and companies that are getting incubated is machine learning. And we're seeing machine learning being applied across multiple projects and multiple technologies, and self-driving cars is one of them. So I'll get into a little bit of those, but this is a reality that we all acknowledge, we all know, it's happening in the consumer world. The question is, how do enterprises take advantage of this? So obviously, this is generating digital natives, as we call them, and uberization has become such a common word now that it's almost too funny to talk about it, but the reality is companies like Uber, they don't own any assets, okay, they don't have any cars, and yet they're competing with the taxi industry. So those are the digital natives. Here are some other examples you may not have heard of. Airbnb, of course, same thing, there's no hotels. All they do is, you know, they have all their assets in the digital world, and they're providing these services. Have you heard of Turo? Anyone heard of Turo? So this is another example. Airbnb and Uber are pretty straightforward examples. Many of us know, but Turo is very interesting. In fact, I happened to use Turo about three or four months ago. They are basically disrupting the rental car industry. So if you have a nice car, like an Audi or a Porsche or whatever, you know, BMW, Mercedes-Benz, and you have some spare time, you can actually rent it out to other people that are interested in driving it. I myself use this because, you know, I was in the market for buying a new car, and I go to the dealership and I get a car to drive, but it's not enough. Maybe I get 20 minutes to drive around, talk to the sales guy, and then I was looking around and I found the site and I said, hey, you know, this is interesting, and I can, and, you know, when you go to the site, you'll see all kinds of people basically renting out their cars to other people. It's basically sharing economy, right? So website, I go in there, sign up, I say, hey, I need a car for the weekend, just like a rental agency, but I get a really, really nice car, the car that I'm looking for, that I want to buy, that I can actually, you know, use for a weekend, take it out for a spin, take the family in there and get an idea of what it's all about, and that's exactly what I did. It worked out beautifully. I actually took a Tesla out for a spin. It was great. The owner was great. He gave us the car. He explained what it was all about. This company doesn't own cars. It doesn't have any assets. All it's doing is connecting people, people that want to rent cars and people that want to drive them. Another example of the digital economy and it's disrupting the rental car business. Here's another example. As I said, machine learning and artificial intelligence. I'm sure many of you use Salesforce. How many of you use Salesforce? Only one or two? Well, Salesforce, you know what it is. It's a SaaS platform for sales guys, right? So they can track their accounts. What's going on? Most recently, just a month ago, maybe two or three weeks ago, Salesforce actually announced Salesforce Einstein. And this is artificial intelligence and machine learning for, and it's being applied to the CRM system. It's incredible. What it does is it looks at all the things. If I'm a sales guy and I have all this information coming in from varieties of sources, right? There's my email threads. Twitter feeds. There's my CRM data sitting in the system and all the social stuff that's going on. It kind of figures out, based on what's happening in real time, that a customer is looking for your product or asking a question about your product at the place you're in that you're driving, you're a sales guy, you're driving around, and you get a message saying customer XYZ is in your area and he's inquiring about so and so. And of course, it does a bunch of other analytics and also figures out whether this customer is likely to buy or not. And then you can pick up the phone and call. Incredible. I mean, this is the kind of artificial intelligence that's being applied. Again, all digital, right? So this is all happening in the cloud. SaaS companies are doing it. Other digital natives are doing it. But what does this all mean? What does this all mean for you guys or for the enterprises? These brand new companies that are just showing up without any assets and they're building these companies that are disrupting traditional industries. Now, you would be surprised to see the kinds of industries that are getting disrupted. We just saw hotel industry. We saw the taxi industry. We saw the rental car industry. And guess what? There's even watches. You know, smart watches, like this Pebble guys or the Jawbone guys. They have, you know, like the Apple Watch. They have interesting different kinds of watches that they build with special information that comes to your watch. An entire industry just doing smart watches. And you've seen the others. You know, WhatsApp, I'm sure you've heard. And some of these other companies are disrupting service providers. What does this all mean, though? This is what we mean by digital disruption. So today, if you want to go build and go attack a grocery business, you'll see that here. There's groceries. You can actually use blue apron, too. In fact, if you're in the mood to go cook something today, some new kind of dish and you don't have the recipe, it's not like a traditional grocer. You can go to their site and you can basically say, I'm in the mood for eating, you know, I don't know, pasta. And I'm interested in Italian pasta with this and this and this. And, you know, it'll come up with recipes. You pick what recipe you want and you order that. The order goes out, it goes to the grocery store and then it gets delivered to your doorstep within the next, you know, whatever, two hours, three hours. Another very value-added service on top of groceries. I mean, you would be surprised. Safeway is doing some interesting things around, so you can actually, instead of going to the grocery store buying your milk and all the usual stuff that you normally buy, you know, groceries for your week, again, they'll say, look, if you're driving around in the area where Safeway is, you just, through your mobile phone, order what you want and then drive by and your stuff will be waiting for you. You can just pick it up and go. Again, you know, digital technology at work because everything is happening through your phone, in the cloud, you're ordering in the cloud, you show up. The next step, and again, Silicon Valley is leading this, is drones, right? So what's happening in Silicon Valley is a whole bunch of new startups and drones that are starting to do grocery delivery through drones. I mean, I don't know how far that's even feasible. Drones are nice for cameras and stuff like that. There's some nice applications. But I don't know if they can actually pick up five pounds of groceries and drop it at your draw step. Again, I'm just pointing out what kinds of technologies are happening. So this is digital disruption. Aside from what we heard in the keynotes, these are the kinds of things that are driving competition and some companies that are looking to compete in the world. So it's a very powerful movement and it's driven by all these new technologies. How is this all being done? I mean, how do you do this, right? And this is the reason why, and I'm sure you've heard of this and this is information out there, half of the Fortune 500 companies have disappeared just since 2000. And part of the reason is what I just described. So all these new digital natives come in and disrupting the traditional businesses and if they don't keep up, if these new businesses don't come up with a way of competing with those or have a competitive service, they're essentially going out of business. So this is the existential threat for traditional business. No matter what business you're in, you may be in banking, you may be in groceries, you may be in watches, shoes, you never know where the next digital native is going to come and kick you. So this is what we're facing as an industry. So what is this doing to customer expectations? So we are already used to many of the things with our mobile technology in the consumer world. It's increasing customer expectations. So anytime, anywhere, personalized, I just give you the personalized experience, right? I'm driving by and I get a message. It's already started to happen with Google Now and Siri and others where you actually get information that's relevant to you at the time, at the place that you're at. And maybe you've served for something in the past and we're starting to, I'm sure some of you may have subscribed to some feeds like this, is at the time that you need something, you get the information. You may or may not choose to act on it, but it's personalized to you anytime because it's on your mobile. It's quick, it's real time and it's very responsive. You just say, oh, okay now I'm driving past SIF, let me go order my groceries and you click and in the next five minutes you'll get your stuff. And of course, it has to work at scale because this is not one customer, two customers. If it's the entire mobile customer base, you're talking about millions of people. So if you're an enterprise trying to serve that audience, guess what your challenges are, right? This is what it comes down to. In the new world, it's not about large, it's about fast. So how do you deliver your services quickly and, you know, looking into Silicon Valley, we are very familiar with this. It's all about speed and about agility and how we go about doing this. So it's not the big fish that eats the small fish, but it's the fast ones. That is the key driver. So these are the two drivers. I said in the beginning, what are the two drivers that make you successful in the digital world, speed and scale? If you can do these two things and these two things you can do in a couple of different ways. And we've heard in the keynotes about software terms, but using software. And we've heard Jonathan Bryce talk about it. Ultimately, we have heard the term software is eating the world. Open-source software is eating software. Hey, it's all software. Digital service is all about software. So this is what it requires. It requires rapid delivery of software services, which means more software projects, faster iterations per project, more user stories per iteration. These are all software terms. In the software industry, you'll know what these terms mean. DevOps. So how do you deliver if you have a website, massive website serving tens of millions of mobile users using your technology, you know, you're pushing out changes four-square. You guys must have heard about four-square. They push out 400 changes a day. 400 live changes into production. I mean, you may say, that's crazy. I'm not going to do anything like that, but that is a kind of responsiveness that their customers are looking for. So this is the challenge we have. So if you have to do this, rapid delivery of software services, you know, you may be doing cloud, you may be doing mobile, machine intelligence, artificial intelligence, IoT, NFV, whatever your workload is, at the end of the day, it's software. At the end of the day, it's software, and the next question is, how do you deliver software rapidly and at scale? That's the key challenge. To do this, you need an underlying infrastructure. You need a foundation. A foundation of scalable on-demand cloud infrastructure. Otherwise, it's impossible. You can't do it. All these digital natives wouldn't have existed if it were not for the public cloud and the private cloud. So this is the foundation, and this is what OpenStack does. And this is how OpenStack helps you to go compete in this digital-disrupted world. So with that as the foundation, let me talk about what the biggest problem is. So McKinsey has been doing some interesting studies, and I'll get to my study in a second, the four super users. I'm kind of building up the story. So this is what they said. Despite their high-priority, highly-visible multi-year efforts, this is about enterprises, companies that are trying to build cloud programs. Half of the participants, they did a survey of over, I forget the number, about a thousand enterprises that are doing cloud programs. Only five percent of them have been able to move no more than five percent of their workloads to the cloud environment, whether it's public or private. I'll pause for a second for you to absorb that. That is an incredibly disappointing number. If you have 10,000 workloads in your data center, you're basically moving only five percent of that. So there's something terribly wrong with this picture, right? So effectively, these organizations over the next 15 years, I can almost guarantee they'll disappear because of digital disruption. So something's not right. Okay, but again, they go on to say something more. The difference between the leaders and the laggards of this cloud adoption, again, is very staggering. The laggards have migrated, as I said, just five percent, while the cloud-savvy leaders or super users, as I like to call them, have moved more than 50 percent. I mean, look at the incredible gap. Five percent and 50 percent. There's something going on here that accounts for this gap. There's got to be something, right? And that's what we're going to talk about. So we'll talk about first, you know, what is it... So we at Morantis have built a lot of clouds for a lot of our customers. And we survey users. We look at reports. We talk to customers. And what we have found out ourselves is there is a common pattern amongst the laggards. And this is what it comes down to. What's going on with these laggards? You know, and this is what we hear, right? We hear... The first thing we often hear is, oh, you know, we have a cloud program and we started off this nice little project to do cloud computing and OpenStack or whatever it is. It doesn't matter. I'm just giving an example. So this is what we hear. But I ended up with very little usage from our internal customers. We can't scale. There are no users. There are no tenants. What's going on here, right? Or we often hear this, oh, there's no business buy-in. We are stuck when... And we have seen this with OpenStack for the longest time. Some would argue it is still in this state. Which is we are stuck in the proof-of-concept phase or pilot phase. Oh, we are still doing pilots. We're doing beta testing of some toy applications. And we're finding it very difficult to get business people to invest. They're saying, oh, this is a waste of time. Where do you want to go build a private cloud? Just go to Amazon or whatever and do some things there. It's like, okay. And that's what's going on there. And the third and the most important thing is, you know, hey, we cannot treat this as a technology initiative without the company committing to a transformation, right? So there's no top-down initiative. There's no top-down... CIO is saying, hey, go do this. So some guy who is in the garage or maybe one of the... It's sort of bottom-ups in some way. So they'll go download OpenStack. They'll play with it a bit. They'll get a couple of users. And that's it. And nobody is getting any benefits out of that. Of course, OpenStack is still very complex. Some of you, many of you are new to this. But over the last six years come a long way. You can very easily deploy today. But operations and usage is still shaky. You know, upgrades and scaling up and down. And day two operations, as we call it. Day two, meaning lifecycle management. They're still not completely mature. So unless you're like a super user, which we'll get to in a second, I mean, you really need to know how to operate this large-scale data centers. So this is what's going on with the laggards. They're having trouble, you know, just taking off. They start off very... with the right intentions. But, you know, good intentions are paved. You know, with forms, right? So it's not working out for them. So what's going on? So what we said is, okay, this being the case, how did these leaders, cloud-savvy leaders, or super users, as we call them, demonstrate business value? All right? And how did this succeed at very large-scale production? What business metrics did they have? What did they actually achieve? What was their formula for success? Were they doing something differently from the laggards? So this is what the biggest question we had. Not just with OpenStack. Of course, OpenStack was what we cared about, but we were asking a broader question. So we set out to find out. So we said, okay. So we went and partnered with Intel. And we said, look, you know, if you want to get this data, if we, as Morantis, as a vendor, we go and ask, you know, we may get, you know, if we publish it, it'll be perceived as being biased, customers will not tell us anything about it. So we basically instituted a third-party independent group called Catalyst Strategies. They're based in the Bay Area. And we said, hey, you know, why don't you go and do this not survey. This is not like a, you know, go to this website and fill out some forms. Go talk to them. Go talk to actual users and business leaders and stakeholders at large companies that are deploying OpenStack. And let us find out. So tell us what you found out, right? So we said, don't go with the very small POC deployments. Go with the very large ones. So our criteria was anything greater than between 100,000 nodes in production, one-on-one interviews with the actual stakeholders that had a vested interest in a successful business outcome. It's not a toy project. Something had to come out of it. And there was a business initiative that was kicked off. And so we sent them off since they were third-party. Many of the people that they talked to actually gave a lot of information because we said we won't attribute you. So unfortunately, I cannot reveal to you who exactly these companies were because otherwise we wouldn't have gotten the data. So no attribution to any company or individual. In fact, we gave some incentives such as charity donations so they can actually provide us this information. So that was the project. And these were the super users. They were from across multiple countries, multiple industries. You can see Fortune 500 financial services company, media and communications, retailers, telecom, technology companies, et cetera. And you can see the size, right? I mean, we didn't go for small sizes. These are fairly large, significant production environments. And one of them had over 20,000 nodes in production. Now, some of these numbers will pale against what you heard at the keynote. Some of them, like China Mobile, is massive, right? Or CERN, which is running hundreds and thousands of cores. Again, those guys are the exception to the rule, right? They are do-it-yourselfers. They're like in the OpenStack community contributing code. I wouldn't consider them as traditional standard enterprise. They're not. Let's face it, right? I mean, I'm talking about enterprises that are out there doing business and are facing this digital disruption threat. How do they cope up with this? So pretty nice, you know, list of people. So having done all of that, what did we find? So I'm going to share with you a few things. First, I'll share the actual business results that they got in terms of agility, cost savings, what was before and after. And then I'll share with you the... So not only did we ask them what your metrics were, how much did you save, but we also asked them how did they achieve success? How did it work for you? What was the journey like? And of course, there's going to be a very detailed white paper that we are going to publish with details because we don't have time today to go through all that. But I'll share with you the high-level summary. What we found out, which is not surprising, I think many of you would kind of agree with this faster time to market. Yeah, we've heard this many times. So what's so great about it? Lower infrastructure costs, more flexibility, and if you go to the openstack.org user case studies, you'll find, I think, more than 300, quote-unquote, case studies with lots of data. What we didn't find was concrete. So everybody said, yeah, yeah, we saved money. Okay, it's very fast. Okay, by how much? How much did you save? Exactly. And it turns out that the laggards, as it turns out, weren't tracking their stuff. So it's all anecdotal, right? I talked to one of the guys and said, yeah, sure, we saved, now we're able to provision in five minutes. Okay, how much was it before? No idea. Okay, our velocity has improved. We're doing more software development. Okay, by how much? So you don't get the concrete data that you need. So all this is fine. I guess we all agree that we should expect to find something like this. But what we found was concrete numbers. Let me share that with you. So here's the key agility success metrics, right? So the first one is reducing cycle time. So if you want to provision infrastructure, and you're working in a traditional hardware world, you ought to go, you know, get machines, you order from Dell, HP, Cisco, whatever, and they show up after six weeks, and then you put them in a data center, rack and stack them, provision them, install them. That's what it used to take, five to six months, bare metal. And these aren't real numbers from real customers, right? Five to six months. So if I wanted to start a project, I can't even start for five months. Let alone get started. How can I be agile if that's slow, right? And with VMs, with VMware, that was cut down over the last decade to, you know, somewhere between half a month to three months. You know, VMware has done a great job in terms of reducing that provisioning time. But that wasn't enough. In the new digital-disrupted world, it has to be faster. So what did we get? The new color is before, they did open stack, and the red color is after they did open stack. So now everything is in minutes, right? You saw examples of self-service provisioning. You just log into Horizon dashboard, you click what you want, and your VM is there in seconds, in minutes, sometimes. With containers, it's seconds, milliseconds in some case. So, I mean, you can see the huge variation in this. So this is not a surprise, but they're real numbers from real companies, and these are published in the white paper. So that's one. The second one is about user stories. I don't know how many of you are familiar with user stories. This is a software development term. So if you're building new features, you build them as a user story, right? Like you say, okay, user wants to do XYZ, this is what it looks like, and you kind of track them based on how many user stories you're actually delivering into production. That's the next one. That's how you measure developer productivity. As I said, it's all about software. At the end of the day, you have to deliver software. How fast do you deliver software? So developer productivity. So again, you see the baseline there. With OpenStack, you get about 20% to 30% improvement in developer productivity, a number of user stories per release. And we have seen some incredible numbers. We have seen up to 60% increase in some cases. And of course, improved time to market. Again, how fast? It's twice as fast and four times as fast in some cases. So now McKinsey did a survey too, and they said developer productivity increases somewhere between four and six. So our numbers kind of fall right into the middle of that. So we know that we are in the ballpark. It's not like these people are claiming something that's not true, it is true. So it kind of fit into that paradigm. So one is agility, which I said is very, very important. And obviously, when you're doing things at scale, you want to reduce cost and increase your IT efficiency. So cost matters. In fact, if you look at all the surveys of OpenStack over the last several years, that was the number one thing that they wanted. They wanted to improve cost. Again, how much by how much. So your cost and your efficiency really depends on how much you're utilizing your hardware, right? So again, before OpenStack, before any of this stuff, I'm sure you have seen this. If you're measuring your utilization, it's less than 10% typically. And in our case, with all of our super users, it was less than 4%. Which is, can you think about that? I mean, you've got all these servers sitting around and they're being only used for 4% of the time, 4% of its capacity. That's gross underutilization, obviously. But once we had OpenStack and all the pools were consolidated and they're all being utilized across the board by various business units, that number went up to 60%. Huge improvement again. 8 times improvement in utilization. VMs, now many people don't actually measure how much the cost is for their VMs per month. Fortunately for our users, they were actually tracking this before and after. And you can see here, in this case, I'm sure there are different numbers out there. You may have your own, your mileage varies as they say. But here, these guys were seeing somewhere at the order of $600 per VM per month. That was their cost before they put OpenStack into production. And after they did, it's 30 bucks. I mean, you can see there's a 95%, it's like amazing improvement in cost efficiencies. And finally, I think the most interesting one is this last one, this is a specific, I think this is the media and telecom company that gave us this example. They had multiple projects, software projects, like more than dozens of them. And each project had their own system operation, system IT type of people. And across all of their projects, they had 100 people. I mean, this is where if you look at, if any of you have done financial stuff, you will see that 70% of your cost is people. And 100 people, they needed 100 people to do system operations before because they were spread out and they were across multiple projects. But after they implemented OpenStack, it was down to two. Another incredible improvement. Of course, I mean, those people might were repurposed for other things or they will let go, whatever. But effectively, it's automation, right? You're automating everything and we're moving into this industrial automation with web and all that stuff. So again, tremendous improvements in cost. So these are just some examples. Our white paper that we'll be publishing shortly will have the actual case studies. Anonymous, right? And that the user was facing. What is the size? And all that is being published as we speak. So that will come out in the next few weeks. So this is a highlight and a summary of all of that. So the next question is, okay, great, I'm getting all this stuff. Why should I even go build an OpenStack cloud? Why can't I just go to AWS or some of the public? You may have your own reasons why you want a private cloud for compliance reasons or privacy or whatever. There are several good reasons to build private clouds. But you may say, hey, I have no resources. I can't build this thing. I'm just going to AWS. Is it going to save me any money? As it turns out, and I'm sure there were a few other sessions today, I think, the day before yesterday. And there's enough data out there that proves that you can actually save a lot of money in terms of cost if you build your own cloud versus AWS. And we did our own studies. We've got OCO savings, OpenStack versus AWS. And again, I can't share the names of the companies here. I think Tube Mogul is the one that did the other presentation. So if you get a chance you check out their recording, they claim about 30% savings versus AWS. And these are the other three our customers, Marantiz customers, that in average have saved about 50%. It's a very, we have a very good model. There's a demo on merantis.com where you can go and compare the costs. It's a very simple online calculator that you can just compare your costs against. It's TCO, it's basically a three-year calculation of your costs over several parameters. Now, this is a simple model online. You can click, you can kind of say, I need 10 VMs and I have this much storage capacity and it'll do the math and give you the number. But there's a much more deeper model behind it. And of course, it can contact us if you're interested. And these slides will be available for you guys to go click on that stuff. But you can go to merantis.com, search for TCO. And we have a model that takes everything into account, right? The number of people that you use, power, cooling, utilization. It's a pretty comprehensive model. We have built over the last year and a half and we have actually tested it with actual customers. And they plugged in their numbers and they gave us inputs. And so, I mean, I think the takeaway from the TCO study, if I were to summarize it in one word, it's utilization. If you can get your utilization to 70, 80, 90%, your private cloud will blow anything out of the water in terms of cost. So that's the takeaway. But you can play with the numbers and see what it is. So with that, I think we're just about time, I think. So I don't know if you have time, questions, yeah. So I'll take a couple of minutes because I think the last couple of slides is important. So how did they do this? How did they succeed? What were the success factors? You saw what did not help the laggards and what did these guys do right? So first of all, they considered this whole cloud initiative as a strategic business imperative with broad support. So the problem is, for the most part, the laggards did very siloed, departmental level implementation, and that was it. That's why it died there. Nobody used it. And when I say broad support, I mean finance, legal, not just IT, security. Because the way you consume cloud is different. Your finance is not used to paying by the drink, so to speak. So there's differences there. There's security, there's legal implications. So you have to bring everybody together. And we have seen all of these cloud leaders and super users did that. They started off by getting a very broad support. And another thing that was interesting is they took, especially some of these large guys, basically set cloud first. That was the imperative, right from the top down. They said, if you want to do anything else, it's by exception. You will not be given resources. You will not be given any support if you're doing anything that's not cloud. And that was it. And everybody was like, OK, well, sounds like I have to do cloud. So they had no other choice. So that helped. And the third one that's extremely important, and the reason why we got such all these great numbers, was these guys were actually tracking. They had metrics program in place. Of course, they made a business case up front with their business leaders and got buy-in. But they also had the metrics in place. And they were tracking these metrics along the way. Otherwise, we would never know before and after. There's no way we could have gotten that numbers. So they had that. And of course, they also, after they did that, but how they started off with low-hanging fruits. So they said, OK, what are the apps that best require agility and work with the teams that get it? Because otherwise, you won't get support. If they're stopping you and putting obstacles all along the way, nothing's going to happen. You need people to support you along the way. So you start with there. You roll it out. And then use that experience and success to roll it out more broadly. So that was the second thing that they did. And then, of course, OpenStack, specifically around OpenStack, they leverage the OpenStack benefits of the community. And they can come to the community for questions. They can go to meetups. There's a huge resource pool right here. Thousands of developers. So many of them took advantage of that. And they were able to fix some problems they couldn't have otherwise. Like I said earlier, you have to educate. And it's a cultural transformation at the end of the day. It's not just IT. It impacts business. It impacts all the different departments. And it impacts competition. It impacts your viability in the business. So if everybody gets that top to down, then it works. Otherwise, it's going to be a problem. So those are the things we found out. So if you have any further questions, I mean, I'm sure we have a few minutes left. So if you have any specific questions about this, again, all the details will be there in a separate white paper. Or do you have any questions right now? Any quick ones? Yes? Yes? Yeah. It's actually, if you look at the cloud. So there is a metric called servers to admin ratio. And this is tracked pretty regularly in the industry. So if you look at Amazon or Google or Microsoft, any of these guys, I think the last number, our numbers are dated, but they usually have numbers in the order of 700 to 1,000 servers per admin. Because everything is automated, right? They don't need sysadmins to go fix individual servers, update patches, the pets versus cattle thing, right? So they don't have, you don't need people. Now in the traditional enterprise, just to give you a number to compare against, traditional enterprises, traditional IT, it's between 5 and 20. So you can see the ratio already. 5 and 20, traditional and 1,000, there's a ratio of 30 or 40 right there. So automation is key. And obviously, OpenStack is self-service. So you don't need, and plus, it's all cloud native. So if a server goes down, you spin up another one and you're done, right? So those numbers are real. And we see this in the public cloud. And we saw them in the OpenStack environment as well. It's usually about 250, 300. Again, it depends on the workload. If you're moving legacy apps, then they require some caring and feeding. If it's classic cloud native apps, it's a different thing. It depends on your SLA. The number of factors, but we do do things like design and architecture sessions with our customers where we work through those kinds of things. OK, other questions? We have just a couple of minutes left. Yes? Absolutely. Yeah, scale is not so, yeah, you're absolutely right. So efficiency, utilization efficiency, and scale. Those are the two factors. And depending on, so even DIY, if you're a do-it-yourselfer, you're not using an OpenStack distribution or any third-party thing. Even those companies, but the scale is massive. I don't have the exact numbers. I have the spreadsheet with all the details. It's somewhere in the thousands of VMs range. Before you break even with AWS, you need to have a certain utilization at a certain scale before you break even and start to achieve the benefits. Otherwise, at low scale, if you're just starting off and you just need 20 VMs, you're a startup, why would I go do all of this? I'll just go to Amazon and I'll keep doing it. And there are many companies that started off that way. They go to Amazon, because it's easy. I don't need to invest any money. There's no capital investment required. There's nothing, so I just go, start off, build my app. And then after you reach a scale, then you take it back in-house. So yeah, we have seen that too. And the models have that, 1,000 to 2,000. Yeah, about 1,000, I would say. I've seen about 1,000, yeah. And then utilization is, it needs to be at least 60% before you start to see some benefits. If it's a very small percentage, like 20%, then it's the same old world. You're back to your legacy world. You're not really achieving any benefits, okay? Last question. Okay, thank you very much for your time. Appreciate it.