 All right, well, thank you, everyone. Can you hear me? All right. Thank you, everyone, for joining. We've got a great panel today. I'm really excited to introduce some of the folks here. I'm just going to name who's here, and then I'm going to let them do a quick intro. So I've got Dan from Getty, Sebastian from Scaler, Shen from Rancher. Rancher, Io, and Mark from Redapt. I'm Jeff Dickey from Redapt as well. But let's kind of kick this off. Can you do a quick intro, Dan? Like Jeff said, Dan with Getty Images, we are the company behind a lot of the imagery you see every day in the news or in editorial content or in a lot of creative stuff. We are responsible for most of the editorial imagery that's coming out to both to websites or web pages as well as a lot of stuff in press. I manage the tech services group at Getty, which is responsible for pretty much everything outside of Dev. My name is Sebastian Seidel. I'm the founder of Scaler. We do policy and governance for a multi-cloud open stack, Amazon Azure, Google Compute Engine. And yeah. Yeah, I'm Shen Liang. I'm the co-founder and CEO of Rancher Labs. We build a Docker container management system. And really, the vision behind it is you should make technologies like Docker should be able to run on top of any cloud and make porting workloads between one cloud easier to another cloud. And ultimately give user more choices between different kind of cloud providers. And I'm Mark Williams, CTO for Redapt. Redapt is a systems integrator that helps companies build out data center scale infrastructure and provides cloud solutions on top of that infrastructure. In a prior life, I was actually using Redapt as my vendor. My prior life, I was responsible for operations at infrastructure operations at Zynga where we exploded into Amazon and ultimately built a large private cloud to deal with some economic issues and performance issues. I was also a customer of this gentleman when he was also the founder of cloud.com, which was the originator of cloud stack. Yeah, so I think in putting this panel together, I mean, we've got some interesting dynamics here. I've got customers, partners, partners that have worked with customers, former customers that have worked with all the partners. And now, you know, part of our company. And we've gone through a lot of different financial modeling, financial issues, financial barriers. But I mean, kind of the gamut. So we're going to kind of talk about that. And so let's kind of kick it off. Well, I mean, first, open stack is free, right? Totally. All day. We all agree that that's free. Why is cost everything in the whole cloud ecosystem? Well, to me, it comes down to IT and the perception or the expectation of the business to have that be a small fraction of a percent of the cost of the company as possible. That derives necessities like laying off people when there are downturns and having all the innovation that every customer asks for at no incremental cost. So just the always impossible problem. It's just inescapable. I think you couple that also with the perception that we are on a daily basis. We're getting screwed by from an IT perspective. The expectation that when you see an advertisement for AWS or you hear about an advertisement from Rackspace or GCE or whoever, the expectation is that they are faster, they are more nimble, and they are cheaper than what your IT organization can provide. So as an IT group, you're constantly struggling with the perception that has already been set with your CIO or worse yet with your COO, your CFO, and your CEO. The one possible differentiator, and this was actually true for what we were able to do at Zynga is the technology teams at Zynga did make Zynga more interesting as a company, because of what we were able to say about what we were doing with Private Cloud. There was technical credibility that added value to the company, and ultimately in the way we proceeded down our path to actually save the company meaningful amounts of money. But yes, you are always compared with the things that are marketed so well, Opistack, that ultimately aren't free at the end of the day and cause extraordinary amounts of pain. So let's go back a little bit, just a year or so, back to the pricing cloud wars. What has really done to the Private Cloud landscape and especially that's really the Opistack landscape. What's happened? Is it helping, hurting, what are we seeing now? So I'll take that. So one thing that's been that I've observed in like maybe a little bit more than the last year, but basically like for the longest time, Amazon was the only game in town, and what we've seen here recently is that because of alternatives on Google, Soon Azure, OpenStack as well, CloudStack before that, your CFO can now have a legitimate vendor leverage strategy where you can source your infrastructure from two or more vendors, so basically you can get optimal pricing. And it used to be you couldn't do that at all just because Amazon was so damn ahead of everybody else that your developers went straight to them and couldn't really move elsewhere. And so now you have a lot more pricing power. So I think that was kind of like the impetus for a lot of the price wars that led on to all of that. I'll add kind of one of the themes I've experienced in my two years at Redapt. So when I first joined Redapt, there were a lot of customers that were emulating that same desire as the Zynga story as moving out of Amazon, which was comparatively more expensive four or five years ago and seeking to have that same story as Zynga did where investment in CapEx delivers long-term savings and a reasonably achievable in terms of time ROI. What happened after April 1st of last year when Google really dropped those prices is a lot of that ROI, certainly the time to ROI extended out and the amount of ROI eroded too. So where I used to have customers coming to me with their Amazon utilization reports saying, well, what does this cost to do as a private cloud solution? Which we could do for them. In fact, one customer I delivered a report to them in March and they were still kind of absorbing that and then April 1st hit, I'm like, I've got to tell them, I've got to do this analysis again and I redid all of the ROI analysis and I'm like, hey guys, I'm redoing this for you because the whole story just changed and no longer are we recommending that you buy a private cloud solution from us, you should stay where you are. The other thing that I've observed is the perception of IT security fears or governance fears about public clouds being safe has eroded quite a bit for two reasons. One is, gosh, when you make something cheap enough, you have less of an argument to push back. The second one is the lack of breaches. You just don't hear, it's not Amazon getting breached. It's nothing, no bad news is kind of good news in the end. I do want to follow up though with your point around stuff being more expensive or less expensive. I think that there is very clear economies of scale that come into play and you see this most obviously in places like Rackspace. Rackspace obviously does not have the scale that AWS does but Rackspace is still making a very good living off of providing cloud services as well as cloud solutions as well as the services that they wrap that in. So there's a question of scale. At what point does it make sense for Rackspace to actually run their cloud on AWS? That's an interesting question for you to ask because obviously it's not at whatever point they're at now. So you keep stepping that back and saying at what point is some type of enterprise customer who's running say 10,000 servers, 10,000 nodes, at what point is their cost to actually run that and manage that internally versus in an external solution? At what point does that kind of hit a delta and do they get some type of strategic advantage like you talked about either in the industry that their company can use? There are places where that economic benefit of using AWS or using other people still even today is still better to bring it in internal. If I can refine that just a little bit. Also the scale question break down at multiple layers like at what scale do you need your own data center? Do you need your own power and cooling? That's a scale that's much further than what scale do you need your own couple of racks in an aisle? So I actually feel like I know clearly lost April, the latest price cut of last April and I think yesterday there was another round of price cut from Google. But I think these price cuts certainly changed the conversation quite a bit and it seems like a lot of people like Mark said will no longer do private cloud for cost saving reasons. But I haven't been through I've been in this industry since 2009 we started really nobody at that time really believed Amazon would be as powerful as they are today and a lot of the best will put on private cloud. And I think there was a bit of inflated expectation for public cloud and for private cloud. And now I think I feel like the prevailing wisdom is it's gone completely the other way. Not only public cloud is the way to go but also Amazon or maybe Google or maybe Azure would take over the world. And I think that's I think we definitely have a tendency to just oversimplify that and knowing the varying degrees of desire for people to consume cloud services. I mean I actually gotten the same group that actually put up the report a lot of people quote. I mean they put it succinctly cloud is by no means a commodity. So there's just so many other reasons that people would use to pick their cloud provider. I have to say I mean one obvious since we go outside of US the picture is just entirely different. In most cases places around the world Amazon isn't even the number one or number two cloud provider. And then we'll find it hard to believe except these countries actually have more computing is gonna have more computing resources than the US. So I think you really gotta keep that in mind. Keep it in perspective. I have a feeling that we're at a distinct point in time. Sometimes down the road looking back we'll realize perhaps this vision of just three dominant players supplying all the computing power to the world is probably not gonna happen. Okay I'll do just one more question. I'll open it to the audience. So we're talking about economics and specifically AWS and the public cloud. If that ROI really makes sense and it's really hard to compete technically with what they have and the speed at what you can do with what they have why are we seeing so much popularity in the private cloud and the open stack deployments and why is that happening? So I'll take that first. So there's a session that I kind of participated in before where NASA was talking about why they're building out their open stack cloud is basically at the end of the day it's all about the data. What they do is they beam about 65 terabytes of data per day from outer space. And that's collected in a data center where they have the deep space network in Alaska. And the pipes are just not big enough to be able to throw that all the way to Amazon. So the only thing that can do is build out a big data center out in Alaska and do the compute close to where the storage is. So there's lots of use cases like that. There's lots of, like in Germany you might wanna have your own open stack there for the privacy laws that they have. So compliance or data are really the big drivers that I'm seeing. Data is certainly not one of them anymore. Yeah I'll add that performance specific requirements for certain workloads is another driver. So for example that customer I mentioned that I said not a good idea for you to pivot to private cloud right now. They were already in a public cloud. They knew their existing performance. They knew what the private cloud would look like but it didn't make sense. But often times companies have workloads that just aren't cost effective or performance in an Amazon or another one. It could be an IO constraint. It could be the data gravity problem. It could be any number of things. The other reasons I would agree is jurisdictional laws that gambling casinos for example, they need to drop what little open stacks into casino regions where it's legal to have such gaming. And you touched on it a little bit. I think that expanding on that idea there's also migration cost. There's a lot of systems that may take a significant effort to move over to an external cloud provider that may be easier to keep internal because of latency challenges because of other challenges like that. But I think also that we need to not totally fool ourselves. There is a lot of hype too. The same emotion and or whatever you wanna put there. The same thing inside of people that says I love open source and I don't like Microsoft Windows is the same thing that is making people say I want to build an open stack environment. And yeah, AWS exists but that's for the rookies. Also I think for a lot of consumers and especially businesses at the end of the day it's never one size fit all price isn't everything. If that were the case nobody would have bought an Apple product or shopped at Whole Foods or they wouldn't be any Herman Miller and everyone would be, all the offices would have IKEA furniture. So it just, the world of just a really low price. I don't know what you're trying to say there. Play supplying all the computing needs while the computing resources is plentiful is not necessarily gonna be the future because at the end of the day unless you are a, you are the kind of business that spend this proportional amount of your revenue on computing resources where absolute 5% saving of computing cost is essential. Like for most businesses in the world I mean they want the good stuff. They want customized stuff. So I actually think there's quite a bit of a bright future for the kind of technology we're talking about here. Yeah, so there's sometimes there's like on Amazon there's like very good reasons to be on it but sometimes there's also artificial reasons. For example, if you're deploying, if you're using Docker Hub, Amazon very actively throttles bandwidth from Docker Hub to external cloud providers. So if you're trying to launch to some containers on Google Compute Engine and you need to pull containers that are no hundreds of megabytes or heavier, they know that's gonna take forever to deploy. So sometimes there's those artificial reasons as well. Is there, what about the human element? Don't we just kind of want it to be ours? Our racks, our build. You want to hug your cloud, Jeff? I do want to hug my cloud. Let's open it up, do you have any questions so far? Can you mind using the mic up here? It's just right, yeah. You guys hear me? Yep, yep. Okay, so the thing I would add to that as a major motivator is just control because the reason to go private cloud is simply because you can control it better. So if you have a mission critical app, for example, that is gonna slow down or shut down your business and you're debating whether to put it on AWS, you look at the risk of failure on that and you could be bringing down your business for 24, 48 or whatever many hours and you have no control over that. Your recourse is to get a refund for your fees. So even though the risk of that is very small, the stakes are too high and if you're sitting there as a CIO you're not gonna take that risk because it's your job. I can tell you I've been through five major Amazon outages and three private cloud outages. They, there's no guaranteed safety by controlling it yourself. Outages are a reality and Amazon isn't gonna give you a credit if they're whole data center. In fact, I think two of them can go down in a region and you're still gonna get zero credit and initially as a consumer of that when it was early for us, it's aggravating. That is not the expectation, but it actually forced Zynga to be smarter about how the applications were designed, how we would deploy them. We kind of, we were beaten into submission especially because while we were early in Amazon, they were having data center outages. They were growing too fast and as a result losing entire power for data centers and so it forced us to be smarter about that and to build in resiliency and how we would do our operations. Yeah, but expand on that. I mean, I think that anybody who doesn't build in resiliency is completely going against the entire objective of what we're trying to accomplish. When we talk about a cloud native or cloud aware application, we aren't just saying, hey, let's go do open stack and it's awesome. We're talking about, it doesn't have to be immutability but it is the intelligence around I am able to understand that my environment won't always persist or exist. Like you're moving away from, and this is not a knock on VMware, this is just a reality of how we use it. You're moving away from a situation where you have a VM that goes down, that server goes down and before we had the ability to do Vemotion or storage Vemotions, like you had to get that VM back up. We need to move away from that. To your question around the control, I have about 120 people, engineers on my team, that I would not literally die for but I would figure to really die for because I know they're that good and I'm so proud of what they've done both in the open stack for our green field and in the brown field. However, I'm not gonna compare them against the massive cadre of engineers at Amazon. I will trust the Amazon engineers before I trust my own to keep a data center up, especially when we're talking multiple AZs and just because of scale. Yeah, and just visibility to what is changing. So yes, Amazon is going to be very obscure about specifically what's changing. They're not gonna give you any visibility to specific maintenance events other than at a micro level, like this instance is gonna die, blah, blah, blah. But it just forces you to build the right instrumentation to know what is changing. What are your humans doing? Because humans are the biggest cause of all of this anyway. Yeah, outages is that what that is? So kind of moving back towards more on the private cloud side in the open stack. I think we have another question. Oh, another question. Oh, I'm sorry. Oh, perfect. Go ahead. Well, I was gonna change the subject a little bit if you don't want to head it and add on. But my question is more about, you're doing a lot of comparison between public and private cloud. But I'm wondering if you've also done the ROI analysis within the private cloud itself, like between open stack and other private cloud technologies. And also, you mentioned that, okay, so open stack is free. We all know that it's not completely free because you're putting a lot of labor in if you're using a pure open source. But there are also products available, commercialized versions of open stack as well. So I'm wondering if you've kind of factored that in terms of private cloud into ROI analysis. That's perfect. I was just gonna ask about, build or buy, like what the different experiences are and when do you do that? When is it DIY? When is it vendor related? When is it open? When is it supported, open? You know, when? I still think cloud stack is a very legitimate option out there. It's very good. It's definitely not as customizable as open stack and it's, you know, customizability of open stack is awesome. But yeah, I still think that cloud stack is, if you fit within what cloud stack offers, very clear path. Yeah, there are other, I mean, I'll keep going on that list. There's Joiant. Joiant is another one that they offer a public cloud and the identical software for a private cloud. They have a great story. But again, it's a very utilitarian offering. It's compute and very basic storage. And there's actually huge value in kind of committing to only subscribing to those things. The more simple the layers that you adopt as you architect how you're delivering infrastructure as a service to your customers, the easier it is to operate, the more predictable its behaviors are going to be. And as the complexity evolves on top of that cloud with applications becoming more increasingly cloud native, your job is easier as an operator or as an IT person. I think you're gonna go down a little bit different path. I think the one other thing that we haven't talked a lot about is the strategic for reason for which you're using any cloud solution, whatever it is. So for Getty, when we look at cloud solutions, we don't look at it and say, like, what are we strategically, what is the best technology or what is the best cost? Right, we, well, I shall be honest, it's not totally to my point, but we first look at the people, right? Our people really and our technology knowledge or capability or desire to attack something is really first and foremost probably to where we go. But secondarily, we look at the strategic decision. So for example, it would be like, I think most of us agree that it would be foolish for us to build an expense management system or an HR system. In this day and age, did you seen concur, you've seen Workday, et cetera, you've seen other, and there are lots of other paths, or sorry, SaaS based solutions like that that would say, that's not strategic to what it is that Getty's trying to deliver. So let's outsource that to a quote unquote cloud provider that isn't open stack or something that we build ourselves, but is very aligned with where we're trying to go. Then for those things that are strategic, that we are building and investing in, then we look at the solutions internally that we use. And I'll be totally honest, as like, and this is going to be almost an embarrassing statement for a VP to make, but we don't look really hard at cost. We look at cost boundaries or cost kind of guardrails. If you're way outside of supportability costs, operational management costs, or acquisition costs, of course, some of those are just going to naturally ebb out, but once you get into some type of threshold, you're going to make a decision based on other factors, other than cost. To play on the people aspect and the strategy aspect, expecting your people to be able to absorb or kind of acquire domain expertise in many different technologies so that you can play the vendor game pit multiple private cloud vendors or pub to public cloud or all three different. That's extraordinarily difficult in terms of investment. It does give you that long term kind of price power and pushback the same way you would have in a Dell versus HP kind of traditional hardware play, but getting your people to really be able to manage all those differences that matter. At this stage of the game, if you're able to build out multiple clouds and test them against each other and run workloads on them, then you are flush with cash and I want to come work there. And we answered the question too about just kind of the cost comparison of each one. Like, is it kind of like VMware more expensive than OpenStack and vice versa? We've done that. I mean, yeah, we've seen that and we have put together OpenStack clouds that are way more expensive than VMware. It's just because of the requirements, but it was that agility that they wanted and it met certain requirements. So, and again, it was more of a talent play like you're saying. And that's kind of an economic decision with your talent, so it kind of goes right into that. We got another question. Yeah. So, change of subject a little bit. Can you hear me? Yeah. Okay. So you made some comments about security and compliance and performance and those things and I think I get the security issue that the public clouds are very secure. The compliance issue is something that I just wanted to get your insight into, do you see any movement from the public cloud providers or maybe even their customers trying to change the compliance type of laws, lobbying for those types of things? I mean, is that argument for a private cloud gonna go away? I'll tackle that on the Amazon front. So right now, Amazon's the only one that offers something like that cloud that's ITAR compliant. It's called GovCloud. Azure doesn't have something like that yet. Google computer engine doesn't have that yet. So that's kind of like the only option you have there and if that doesn't fit what you need then open sex the way you need to go. So that's on the ITAR side. On the HIPAA side, Amazon is technically HIPAA compliant but the problem is that they won't sign the BAA which means that they won't assume the liability of data loss that HIPAA forces onto you. So if you lose some healthcare records or things like that, Amazon won't take the liability for that and that's natural because it's self-service and which means that you can screw yourself up and Amazon doesn't want to just eat that liability for no cost. So right now, so if you're building out a cloud that needs to be HIPAA compliant and you need to have that BAA sign with your provider open stack really is the only way to go. More private cloud really is the only way to go. There's lots of vendors out there of managed cloud solutions that will want to sign that for you or with you and that's because they'll do that architectural review and vetting with you so that the system is truly cloud native, you can't lose that data or there's lots of safeguards in place therefore they're willing to take on that liability. I also wanna quickly mention something, this stuff really very sorry, this stuff really varies country by country as I said earlier, we've seen that sort of decision really impacting ultimately what technology they choose and I think the good thing about an open source solution like open stack is basically you got a lot of flexibility and you can operate it yourself or you can easily get a system integrated or operate it for you. I was gonna ask you, Sebastian said kind of open stack is that way to go for security on that private side? Is it, is open stack the option for security and compliance? Well, I'm trying to keep my own opinion out of that but that's just like I've just seen it over and over again many, many large customers doing no credit card processing, managing healthcare records, managing very, very sensitive intelligence agency information and all of that is they all choose to have a tier in which certain workloads are placed in the private cloud and then there's tiers in which it's less sensitive workloads are placed in public cloud and being able to have a portfolio of placement options for your workloads is very important especially the larger the organization the more diversified the applications and workloads you have, the more important that becomes. One quick thing I think security is more of a process than a technology so it's all about that discipline about the vigilance and the procedures and standards where it is important but I think it's secondary. You know also like really in terms of technology I wouldn't right now I wouldn't equate open stack with say private cloud. You know I think when a lot of people talk about private cloud these days unfortunately what they mean is a v-sphere cluster and that is I think the meaning of private cloud has evolved quite a bit in the last few years and so that's why I really hope that technologies like open stack, cloud stack continue to grow in the private cloud space because I think they do have a significant value add to some of the other proprietary technologies. I'm not trying to pimp VMware here but let's be careful that we're not becoming bigots with that statement though because remember the purpose of what we're trying to do from a cloud perspective private or public is not about the technology it's about the business enablement side of that technology is actually a business partner not just a ticket taker. Whether or not VMware is the right cloud for you or here and I know this is blaspheming saying it's an open stack but that's a business decision as long as you're delivering on the objective around faster time to market with business needs around higher velocity to deliver on whatever the business objective is and or around availability we're winning as a team as an IT organization we're winning in helping our business succeed. So every one of you have you've all been involved in several or many large publicly noted clouds and they you know kind of notably successful clouds. Now what what how do you define success? Naming. Yeah the Z cloud for example it's always naming. Everything started with a Z it wasn't hard to name that at all. You guys have a cloud start to the Z? So I was a team a long before I met Mark and so this is in 2010 2011 at T cloud. Yep yep and then rocked it over at Getty and it is. Yeah you guys can figure out where that goes. What is it all about economics what is that? How do you guys define a successful cloud? Well as the business objectives matter right? Having in operations we were all about availability, cost and deliverability in terms of speed delivery and the battle I was fighting was most of my customers those game studios were in Amazon and they were fine with it and they were happy there and most of the business wasn't yet looking at how costly that was and we had already had raw bare metal servers and data centers running other games and we knew they were way more performant and we knew that we could do a better job than some of the issues we had with reliability and what we predicted to be future cost with games growing there so we sought to go build Z cloud. So what success meant for us was making it better than Amazon and that was very much possible but kind of unpredicted. It was so much better than when the first games moved in we were able to predict before they moved in that it would be a third the price and twice the performance. When they moved in what they also was the gravy on top of that was the user responsiveness. We had richer CPUs in our Z cloud that actually resulted in a more responsive game experience for every user and it wasn't ever predicted that way but that was the gravy and that was kind of a catalyst that once I got my first tenant, which was hard to do, everybody was more and more willing to move across and you needed evangelism along the way to make sure that that could happen. Time back and getting a little more philosophical with it you started with talking about the tenants. I think that the way we measure success on a cloud is the exact same way we measure success in any other area against the metrics by which we run our business. So from my side in my organization we kind of hold a loose framework around what I've kind of created as an AVM model so availability, velocity and maturity. So if you can say like from an availability perspective we are meeting customer expectations kind of it doesn't have to be an SLA we don't have to get archaic or draconian about it but we're meeting expectations with the available and the investment we have there. Tied the couple that and almost for me is more important, we're meeting the velocity expectations of our business. We were talking a little bit ago at a meeting if we, it's on our case it's our marketing team wanting to try stuff out. If we tell our marketing team that they can try one new idea out once every two months that's a radically different proposition for them then you can try out 10 things and nine of them may fail and that's perfectly fine because we did it really quick we had a low cost to get that idea out the business idea out through technology or enabled by technology. So that velocity is really important and then finally of the maturity do you have a clear roadmap or you continue to drive up the value drive down the cost et cetera and so I measure it the exact same way I measure other investments. So I think from our point of view we're mostly a vendor supplying cloud solutions into these organizations, IT organizations and web companies and for us the measure of success is actually very simple it just boils down to all boys down to adoption just how many people use it how many virtual machines are created just like Mark said after he deployed ZCloud once we overcome the technical challenge of getting the cloud deployed then the next challenge is adoption how do we get people bring workload on it so that's just like any consumer services you just start measuring those kind of metrics you start measuring churn you start measuring what kind of problem people are having and then solve them then adoption grow and then the cloud becomes successful. So I'll add a note to that so skaters we've worked with APIs from Azure to Google Compute Engine to RackSpace to pretty much all of them and from really from a technology perspective like a shitty cloud is going to require 10 of the identical same API calls to be able to get that virtual machine started and once it started it's gonna take you 50 minutes for that thing to boot up and once it's booted up it might have intermittent connectivity issues I mean that is a really bad cloud experience which leads to all sorts of frustration from the developers which ties to what Cheng is talking about which is like you're not gonna get internal adoption if that's the cloud you're offering your developers and awesome cloud is gonna be something like Amazon or even better Google Compute Engine where you make that call to Google you're like within 15 to 20 seconds your VM is up and running and you can SSH into it and you can start doing stuff so huge amounts of variability and I've seen OpenStack clouds that where you have like boot from Cinder but when you're trying to boot from Cinder because of some network issues it just takes you like an hour to get that VM to be launched so even within OpenStack there's huge variability in quality so it's like ultimately yeah I'm not sure there's a real conclusion but the quality of the cloud is also measured as like real technology metrics such as like those hard numbers of getting stuff done Let's do a question, go ahead Just got a quick question you know we've obviously OpenStack we say software is free unfortunately a lot of the other software we have to run isn't on top of the thing any experiences with the kind of ugly barriers that software companies put in front of portability of workloads across various flavours of cloud you know Amazon took 18 months to negotiate with Larry Ellison how to get Oracle licensing worked out on that platform as a classic industry example any experiences you've got in that space because that's economic impact you know we're all paying for licences in enterprise agreements and stuff already Well in the case of Amazon versus Oracle that was kind of like the Cold War nuclear negotiations right like both parties you know didn't absolutely need to get it done straight away so they were both playing the waiting game and playing chicken yeah I would say I don't know if I won't give specific examples internally but I would say we have had problems with Oracle we have shockingly had problems with VMware we've had problems with the challenges with I call it HP like some of the some of the vendor it's not so much vendor lock-in it may be maybe there is some element of vendor lock-in but there's also a very real element of it's way easier and way faster to get a product out when you don't have to test it across 30,000 different solutions there are a lot of modules in OpenStack and so from that standpoint there is a lot of we'll say incompatibility maybe it is nefarious and it is vendor lock-in maybe it's not but the number one challenge we've had is is more along the lines of module incompatibility heterogeneous yeah not specifically to your influencing a commercial license kind of threat but other strategies to deal with disproportionately expensive services is you know prove to that vendor that you are starting to differentiate yourself with a competing technology that is open source like my you know demonstrate my sequel in some capacity I don't know if that's going to bother Oracle with as big as they are but you know that concept has worked for for many of the companies I've worked for to get that price negotiated actually I'll say real quick one is not directly related to the question but it's almost hilarious it has been funny to see how differently Oracle treats my sequel and and Oracle proper that there's be user completely different the way their sales team sells against each other is completely different it's it's it's an interesting thing just to watch to your question around watching that commercial viability versus a what was an open source and still has open source expectations being trying to be run by a a large enterprise so in just a kind of a a thirty second answer like being kind of fiscally minded and bringing up your private cloud what I'm going to go with each of you like how do you start what is that what's that kind of golden roadmap a lot of customers want to start with oh I want to start with the cheapest hardware the open source version of things and they don't realize how many problems and how many decisions that they're subjecting themselves to that are going to stall them out I think the measure of success is how quickly can you get something operational so that you can begin to innovate that with how else you already deliver I.T. and services to your customers so starting with turnkey solutions you know appliances are good although we just lost Nebula but that you know we've been working with with uh... marantis to do a turnkey like it's ready to ship it's a lot of things have been predecided for you but that's in in order to save you from playing with the box of razor blades so that you can start and begin to learn about this very complicated technology that in a way that has been predetermined to be safe that there are a lot of known quantities around it so that you can begin and then as you iterate it's important to think of your first step in cloud being one that is going to be the first of many you're going to be evolving the choices that you make in in your future clouds don't start with a completely open box of tinker toys start with something that that has got some form to it some structure to it that you can count on yeah i think you know yesterday's keynote which i really liked a lot yesterday morning's keynote very technical keynote and i think uh... my mark mentioned the mark mentioned the uh... technology now thank you everyone