 You done, Jesse? Well, I don't know what The start time for this is 510. So that kind of tells me we've all got mop-up duty Do we have to all pick up a chair and carry them out and stack them up when we're done? Thank you for joining us. I know it's late the evening. We're probably the only thing between you and dinner So thank you for joining us We're gonna be recording this episode as a speaking in tech podcast episode Which will be available on the register on Wednesday perhaps this week perhaps next week We'll see but I want to start off and go and kind of go right down the line with our panelists This is gonna be kind of an interesting panel on open stack as a service or as a distribution and Caroline will start off with you if you introduce yourself give your title and then we'll move right down the line Hi, my name is Caroline McCrory and I run syndicated research at Giga I'm I'm Manju Ramanathpura. I work for Hitachi data systems I work in the office of technology and planning and I focus on Network solution innovation as well as open source innovation within Hitachi I'm Jesse Proudman founder and CTO of blue box. I work in the office of blue box responsible for delivering open stack as a service In case of any confusion My name is Ken Hui. I used to be at rack space on the open stack team Currently I am an emcee on the cloud solutions business unit focused specifically on open stack Yeah, no young holds a redhead. I'm part of the technology strategy group on redhead Focusing on everything infrastructure deployment of new and emerging technology is working with our customers I'm Boris Ransky. I'm a co-founder and chief marketing officer at Marantis All right. This is the all-star panel. Thank you guys I want to start this off this the topic of conversation came up in and Jesse correctly if I'm wrong it was as a tweet chat Conversation that kind of started it was kind of new to me that that The that were dissecting open stack as a service or as a distribution. It seemed like wow How how evolved has open stack become to now? We're getting into the hard questions about how customers are deploying and leveraging open stack So I want to tee it off with you first with some of your quick thoughts and yeah We're gonna go we're gonna go to you next because I really look at red hat and blue box is kind of the two antagonists here We're all friendly, but let's let if we're gonna compare service or a distribution This is the the right folks to have so Jesse right off the bat. What spurred this conversation for you? Why I mean obviously You can give us some background on blue box and there's a self-interest you have in that But there's a Phil's I know you're philosophical about this as well. So go ahead. Yeah, so I'm a little biased on the topic But I mean quite frankly we think distributions are just doing it wrong We'll throw it out there What what is cloud I mean, I think that's where we need to start and the whole premise of cloud if you could the NIST definition is about giving users in an organization elastic resources and As we looked at the evolution of what it takes to now deliver that type of technology you get a really complicated set of of Software and so open stack itself people always say oh, it's like it's a bag of bits. It's a something you've got to assemble the reality is that the Elements inside of open stack that you must know about to effectively operate an open stack cloud It's an enormous set of requirements. And so what we've seen particularly in our Customer base is that there is a desire to run a private cloud yet. There isn't the capability to operate that private cloud themselves and We launched blue box cloud a year ago now at Hong Kong With the intent of delivering open stack private clouds as a service and we were the first Company to do that as a productized offering and we've seen some pretty great success with it over the last year So much so that we've now got folks like Marantis joining in the fun But again, I think that the key here is Operations teams inside of an enterprise should be focused on building value in their enterprise And that's often Architecting and engineering things that sit on top of the cloud not the cloud itself And so if we can make that easier for those operations teams, then it's a win-win for everybody John Well, I'm biased to I guess But I personally this was make a great platform for open deck as a service provider So clearly I think we have a long history of making open source enterprise consumable and also making it scalable in terms of Accessible to a large number of customers who just want to consume the technology The service side I think is certainly interesting and there's certainly a place for it But I think there are also a lot of challenges around open-stack as a service for a good number of customers and consumers here I think the key and I'm sure Boris has some opinions as well as to make open-stack consumable for customers and provide life cycle around it provide the Capabilities around upgrades for example all the enter or a lot of the enterprise features customers are a custom to Just now running new workloads and also being able not just to deploy open-stack only but Usually many customers deploy much more opens than open-stack open justice Just a tool underneath for them They want to add value on top of it as just it for example running past platforms running the applications And that's really I think the value is providing the entire Portfolio around it being in middleware being it Management capabilities across a very variety of infrastructure being a public being a private cloud because we certainly see a lot of Customers who want to be able to leverage the resources from wherever it makes the most sense to them Which could be on-premise it could be off-premise, so I don't think there's a single sort of it must be all on-premise It must be all off-premise. I would imagine that red hat is kind of like Switzerland here in the respect that you work with You've got Partners that do both right is that fair absolutely absolutely and it's kind of my point is I mean we We are more than happy to work with service providers. We don't want to be a service provider ourselves We are happy to work with service price We have a long history of enabling a lot of the enterprise clouds out there Based on Linux based on retic technology And again being Switzerland because it's great to partner and we don't need to be the one all and be all Except you run OpenShift as a service so you are a service provider I mean, I think I wanted to jump in here real quick and actually step back one second So what is the definition of a service? What is the definition of a distribution because I think that's a key fair That's a key component to this so When we say sir as a service what we mean is that there is an external firm that is actually operating the cloud Blue box cloud means we actually run the hardware It's it's in our data centers on our hardware Meta cloud delivers open stack as a service on customers hardware in their data center So that the the key distinguishing point is that the operational? Responsibility for that cloud itself lies to folks outside of the organization and folks who have a specialized skill set Where a distribution is a set of software that is is shipped or downloaded or installed and the operational burden then falls on the actual Organization that did that installation and so I think That's a pretty important context to have for this this particular discussion Actually, would anybody disagree with that definition we go to Boris real quick They you look like you're jumping out to speak before I'm so I'm getting yeah, yeah, you're you're a little shy We're avoiding I'm a shy guy so Well, I mean surprisingly I would like to agree with what Jesse said I think that This is like a little bit of vague as a service or as a distribution. I think that as with any Software and open stack in particular there is an infinite spectrum of ways to deliver it where and in that spectrum You can classify where on one side of a spectrum you have a very opinionated way to do it And then on the other side you have the way with least opinion so the example of a way of least opinion would be How morantis started is basically doing services around upstream open stack bits So if you're a large Web to company that is convinced that you know infrastructure Then better than anybody else around and you probably write about it You have a lot of opinion about how you want it done and you just need that fill in the expertise gap To just get you there to you know to to to get running so this is where you get services and then you can move down all the way to the other side of the spectrum which is Rec space public cloud that is using open stack bits But it's not a managed service in their data center or your data center It's pretty much access to an open stack API endpoint for a fixed amount of money and in between these two There is a variety of things so From services you can move to a distribution where you've you know added value on top of the upstream Open stack by hardening it and making sure that it's easier to consume you can go to the next step and say that okay We're gonna Help you operate that in your data center and you can go to the next step and say we're gonna actually do a managed service type thing Where it's our data center You just tell us you know what version open stack you want and how you want it and we'll just give you open stack But it's going to be your rack with open stack running on it And there's an infinite number of kind of permutations in between and all of them have value to different types of customers and different use cases Great. Can you I know you? Wanted to chime in there. Yeah, so first and foremost All apologies to Jesse anyone who tells you there's one best way to do something they wrong That's there's a multiple ways of doing things There's a lot of customers that I would say you need to you should go ahead and hire service provider run your open stack for you you can't do it yourself and you don't want to But there's also a lot of companies who would laugh at the idea that you who They would say right now I have ten times the staff you have that managed ten times the number of servers that you that you currently have in the management Why am I handing the keys off to you to manage my environment? I'm gonna do it myself Jesse that's a fair question. Let's let's arbitrarily say somebody like eBay, right? Why would eBay want to want to work with blue box? Yeah, they wouldn't like eBay would do it themselves And I so I was gonna say to Boris's point there. There's absolutely a big spectrum of options and Quite frankly the the large enterprise players even in me your DMC, right? So you're talking with those large enterprise companies The reality is that the majority of Businesses out there aren't eBay. They aren't PayPal. They aren't Facebook The majority of businesses out there have an IT staff of five to twenty people who are already busy on a day-to-day basis Don't have time to learn a new technology Don't have time to learn the operations of a new technology But need to benefit from private cloud and so when I say distributions are wrong for that type of consumer My goal is to get open stack in the hands of as many people as possible But I absolutely I absolutely agree. I think that the largest organizations Will do it themselves because they have that team so So I think When we talk about enterprises, most of them probably fall somewhere in between. It's not 20 people IT It's probably about 50 people hundred people, right? And they're also not like the source product like You know eBay or AT&T or telcos So I think the the if I just follow your own trajectory, right? Let's assume that You've got like blue box metacloud that are providing Private cloud as a service maybe say metacloud providing an on-prem private cloud There are thousands and thousands of enterprises, right? So there's going to be a point when let's say you're a doctor you will and you conquer the entire world You know, you've got like three four really main You know the open stack service provider for on-prem, right? If you are serving hundreds and hundreds of customers There's going to be a point at which you are going to optimize How your cloud is delivered as a package software? And that's nothing but a distro in its own way, right? You've just created your own version of a distro We were on a panel earlier this morning, right? One of the customer was asking, you know I don't want to we have like four storage vendors emc net app Hitachi data system solid fires and then we had red hat and the guy One of the guy a gentleman came and asked a question Why don't all the rest of the storage vendors provide your storage as a software as opposed to integrated hardware and a software? Now if you take that same analogy as a blue box providing an on-prem private cloud service What you're really doing is a cloud in a box and you're doing it repetitively for hundreds of customers and You've got a software that is pre-packaged like a distro So the next thing that your customer is going to ask is why are you packaging your software with the hardware? Just give me the software instead of the hardware It's just an evolution of we basically took the storage and you know evolved ourself to the cloud You know fast forward maybe five years ten years down the line what you're going to have is basically blue box distro so We have a distro quote-a-quote. It's open on github You can go take our software right now and you can install it on your own infrastructure The customers that come to us specifically don't want that because they don't have the ability to actually operate it Like that's the key is operations right so even if like if I go by HP helium I can still hire HP to manage my installation right that this trend towards Towards external management of the infrastructure. I think it's really critical And it's in my opinion particular to cloud technologies because of the complexity of the software This isn't just my SQL. This isn't a single service like I think that's and in terms of red hat I think that's where I see the greatest challenge, right? Installing a single OS and keeping an OS updated and patched is one thing But now you've got this this massive set of interconnected technologies That becomes a lot more challenging to deal with the failure scenarios on and I Just we've seen a desire for people in organizations to not have to touch that I think that's just the evolution of you know a Linux was pretty batch in in bad shape to when it got started and over a Time it got better. I have a feeling open stack. We as a community need to help for open stack to get to that point Well, I think one one thing here is we need to be careful that as we said earlier There are different market segments as well So I think often the customers you deal with might not necessarily I would almost say customers Abhorus ourselves would necessarily target as the ideal candidate for an on-premise open stack installation because as it they're totally overwhelmed by the complexity We are going to fact that open second shiny object. Everybody believes they need to have at least one because everybody else has one But I was a it's still a box of razor blades to do today touch and install blindfolded So I think the reality is yeah for those people who can't deal with the complexity a Service and I'm keeping service open at this point if it's like a managed service on premise of a customer if it's something like you guys provide But there's certainly a need for those type of customers to to make it as seamless as possible And shield them from a lot of the sharp edges for example So they can run their businesses, but don't necessarily have to deal with and they don't have the staffing to do this So absolutely and just quickly because you mentioned open shift. So funny enough So open shift was just a reverse part for us. So and we are actually not Providing really a managed service we charge for really it's a free developer service But the key for us is we built it to actually Understand gain operation and knowledge of running a large-scale pass service and then customers came to us Wow, I would like to run this on premise So yeah exactly happen and we had to figure out how we actually package this technology to make it consumable for customer on premise without having to have X number of people on staff to figure out how all the Secret source works behind the scenes, but yeah, I think it's a great exercise for us. For example, I actually have real-life Operational experience which often as you deal with vendors not every vendor has actually real operational experience I think you have to know where are the difficult address Thanks, so I wanted to add that I again agree I think that actually creating the software and then operating the software are two different things and two different Kind of components of value add What another thing I think is important to keep in mind is you can look at Kind of historically how the markets around the same things have evolved So the distributions are not new and managed services are not new But the way typically the markets evolve is you typically get a limited number of Software vendors or distributions and then there is operations vendors that do Operations around those dominant platforms. So if you go to rack space managed services and you ask for You know, I want some virtualization rack. They come to you with some Standard solutions. They say you can do Microsoft or you can do VMware and by the way We also have open stack and we'll do managed service for you, right? But It's it's it's the Microsoft and VMware It's the platforms and the distributions that ultimately get the maximum leverage out of their business Whereas operations model typically scales linearly and is more similar to the services business model because ultimately to do Operations you need more people for each additional customer which you get which is a platform You win the mind share and you scale it across the ecosystem and you get other vendors to do operations against your platform I'm gonna go to Caroline real quick. You'll be real quiet here, Caroline. What's I'm waiting. You're waiting the pilots. Yes. Go. Here's your chance. I'm like a giant spider. It just hides I Actually agree a lot with What Jesse Boris and also Ken actually said I think that because the biggest mistake that a lot of the vendor companies Say is their distributions are the only ones that you should look at and they solve all problems And it's absolutely not true every single vendor has written their product for a specific use case And it's actually forcing them to admit what that use case is before you actually choose what's actually best for you same with the service providers and and as Boris had highlighted the way Morantis had started was to do custom and now they've taken a lot of that knowledge and put it into their distribution which is mainly You know run being distributed with fuel and things like that So it's it's very important for you to understand what your business is trying to do what your goals are and Which of these distributions or services has the best? Service level support that you need because you are all accountable for different things to your end customers And they are accountable to you only up to a certain point So it's where is your focus if your focus is your own markets and your own customers Then look to a provider that can deliver the needs that you have to meet your business goals As opposed to looking at it and being mystified by the technology and not realizing what you're using them actually for It's very easy to drink the Kool-Aid cuz you know, everyone's sitting here looking quite cool. I Want to ask Jesse a question Jesse one of the big drivers for private cloud is being able to control of data security of data privacy being Whether you do it poorly or you do it great. In fact, it's all within your grasp I guess one of the challenges that I have wrapped in my mind around Open stack as a service is that you give up some of that control whenever whatever that data is outside of your building How do you how do you address that challenge that companies have when it comes to the security privacy and data control? Absolutely, so I Think I need to define what blue box cloud is first answer that so blue box cloud is a single tenant open stack installation so every customer that we get gets a at least three node Dedicated open stack install that belongs just to them So at any point in time we can walk into our data center with that customer Or with their auditor and say this machine that machine and that machine Comprised where your data is so what we tried to do was blend the benefits of public cloud elasticity time to market op-ex billing and The benefits of private cloud which is data sovereignty control of the environment control of cost and integration to existing environment into into one Unified offering so Because we're not sharing anything we and we don't like there's no dedicated say or there's dedicated sands. There's not a shared sand There's Dedicated firewalls. There's no shared firewalls everything in the environment belongs to that customer In our opinion we can get the same level of isolation that you can get on-premise just without the operational burden But what about is and I'm gonna cherry pick here It may not be fair But if you look at in Carolina, we've had discussions in the past about the Microsoft issue and the data sovereignty Issues that came up with the Ireland Cloud email accounts if if you've got a customer that Let's say the federal government wants to grab their data. We'll assume it's encrypted as part of the service What's to prevent what's to stop them for for asking to see that data? And I know it would be a little bit I'm making a couple of distinctions here if it's on-prem, you know within their four walls and it gives them a little right to at least be alerted to such an action where Conceivably if they're using a service provider, it's outside their four walls They may not even be aware that there's an action to seize that data in the first place I know it's an extreme extreme case and 99.59 of customers are never going to see that type of experience, but what how do you how do you handle that? What's the how do you manage that control of data aspect? Yeah, that's a great question So I mean we have corporate policies in place. We're safe harbor certified So we follow all the legal proceedings that we can to ensure That there are policies in place to prevent that type of action Or to communicate it to customers if it occurs sure we've received subpoenas before and we've had to respond accordingly and Yes, I think that that is a valid concern for being outside outside of a corporate facility But the reality is in my experience, that's primarily fun like At the end of the day majority of organizations aren't gonna have to deal with with that type of Concern and if you have that concern of your data, then you're probably gonna put it on prem Like there are workload. There are use cases for all these these technologies the other thing that's unique is that We've architected the solution so that we can drop into Data centers around the world so instead of having these centralized hyper scale facilities like a public cloud provider does we could theoretically be in 120 250 data centers around the globe and so we can get into more countries So you look at like Amazon building their Frankfurt facility like we could have a German facility We could have we have a Swiss facility right now. We could have a London facility And so by getting into the countries of origin for where that data is you start to get Over some of the challenges of that data sovereignty question Yeah, so I think The old way of thinking about building systems was always where are my servers? Where am I gonna put them? Where are my data centers? And therefore that's where my servers have to be and if we start thinking about things more as a service and Everything is going towards applications. If you understand which applications you have that need to be Internal or external and have data that it needs to access that are internal or external that will help you actually decide an architect Where what you have is a distribution if you have to choose and what you have as a service again Also, if you have to choose it's making the focus back on the applications and the data The applications need to access as well as where the users are coming from of those applications that will help you decide the geo locations of where you have them and then Obviously the rules of the geo location like what Jesse is saying, you know being having something in Frankfurt because there are certain you know legal requirements for the apps that you're going to have there versus somewhere else and then it's also a Very much to do with where the data is going to be located that are going to be accessed because that will change How you also architect your application landscape? Or so as I work work great. How often do you get that question about control of data? Within the sphere especially with open stack You know we do get it from time to time, but I can't say that it's a predominant concern at this point and in all honesty I think that The Model where Open stack is consumed as kind of a you know as a managed platform It's it's definitely going to be something that that that's that's that's a market That's an emerging market But if you were to for example, you know ask crack space right now, you know how much of managed open stack They do versus managed VMware It's you know silly question I think that the main market that we've seen for open stack is among still to date fairly kind of a high-tech kind of a High-tech web sass type organizations that are Adopting open stack with a vision not to replace something necessarily but to solve for The problem that that that open stack is the only solution that can can be solved meaning, you know build the internal Amazon and Most of these projects are still early and most of these projects are always on-prem and the customers always control everything and They answer this question themselves So you know if we when we build an environment for Webex or build environment for PayPal They don't come to us asking us about you know the data security problem They can take out distro and we can help them implement it in a particular way But they ultimately control the entire thing and this is a way to kind of avoid this question altogether Understood. So the title of this panel is ask the expert So if anybody's got any questions in the audience, please look there's two microphones over here stand up to the microphone and We'll jump to it. I can't I know you want or just kind of throw Jesse take your pick Thanks, so I think what we just talked about kind of speaks to the fact that While there it's again that there's definitely certain use cases where you definitely want to use private clouds of service But there's blue box more at this There's also a lot of use cases where you have to do it yourself. And so it's important I think that's where the hybrid cloud comes in right? What whatever you choose you should choose something where you can you can actually be able to Talk from your private on-prem environment to an off-prem environment and be able to put the workload in the right place And I think that's it. That's one aspect that almost all the business have it done wrong have it wrong to this To date which is they don't look at any they don't look at the workloads to examine where which one should go where they just Basically say which saves me the most money in the short term. Let's go to our first question You give us your name if you wouldn't mind. Hi, I'm Raina Moser. I'm currently a software development manager at Rackspace working on the public cloud deployment system So whenever y'all talk about Rackspace, I'm like Really really it's that easy So so for me we we are packaging our own it's it's not a pretty package It's not elegant, but we are packaging code and configs from upstream open stack with our internal changes I know that right now we are we are stuck on a Debian distro and Kind of in a lockstep of how the heck do we upgrade this? We're also kind of stuck in that same problem with Zen server What's to keep distros from making that same pain-point problem because little clouds grow up to be big clouds And so for me where I'm standing a distro just it for for The large-scale use cases that we're going to continue to see with research and development with large-scale data management I have I have concerns about the distro the distro path. And so I'm just curious Once you have I'm biased I am I am biased so I can I can try and answer that question I Guess my question to you is In reply to this I wanted to understand that you work in in the public cloud part of Rackspace or the managed services Part of Rackspace So the managed services is just a new level of support offering the infrastructure is there You either can get public cloud services with basic text support or you can get public cloud services that give you the full But I think you're you're talking about Rackspace private cloud I'm talking about I'm talking about Rackspace managed offering like I can come to Rackspace now and say Rackspace I want to rack with VMware on it. Tell me how much it's gonna cost I am in the public open stack cloud not okay, then the VMware. Okay, so yeah, so the That's that's that's a good that's a good clarification points I'm going to talk about how we deployed a 42,000 things in the public right so when it comes to a certain level of scale I think Distribution is not necessarily a way to go especially if what what what you're doing is Your core business so when it comes to so Rackspace managed service and Rackspace public cloud are two very different types of business So Rackspace public cloud in my opinion is closer to what AWS is and Because of that the open-stack bits that you guys are working on and the entire infrastructure underneath is really part of the core value proposition that Rackspace delivers to the customer and moreover Because it's a public cloud you can actually kind of do the Amazon model where you have a Fairly uniform infrastructure underneath and your own distro that you roll with a lot of resources dedicated specifically to this distro And you know in that case when when you're talking a lot of scale Distributions not a way to go you don't want to go with third-party vendor and this is really the same reason why Open stacks been attractive to a large-scale infrastructure operators like workday or or PayPal Because they don't want a distribution. They want to have their own thing much like you guys roll for the public cloud That's great. Thank you. We've got we have another question here. Oh, I'm sorry. Yeah, go ahead. I probably should chime in too Yes, you should We somehow have a fight and a dog in this fight as well So I think we certainly Have proven that having a distro as a stable base gives you a great platform for building a very large-scale infrastructure because again You get a stable base with all the life cycle rounded updates, etc You don't have to do this yourself. So first of all second of all I think it is really important and we all like to talk about well open stack and OS and whatnot But there's a very tight coupling today Realistically between open stack as a platform the underlying OS a lot of interaction I think you want to make sure that sort of at least the platform you deploy is Going to work and you don't have to sort out interaction between say, I don't know networking Colonel drivers name it namespace support whatever it might be on your own and Effectively not just create your own open stack distro, but also create your own open Your own operating system distro for example So I think the key really is provide first of all infrastructure support around life cycle management deployment, etc But also on a higher level and that's not necessarily direct as Important to Rex because you do it a lot on yourself as you will do the talk on on Wednesday But for a lot of the enterprise customers on premises all the two unify the operational aspect of their different Environments because not every customer actually no customer almost has just open stack in the environment They have a lot of other stuff as far as that like VMware other stuff in the virtualization space and for them I think it's really important to have a unified operational surface for example making provisioning consistent providing workload orchestration across multiple infrastructures rather than just have these silos and Managing each of these silos separately having different skills, etc. But reuse skills as much as possible. So anyway, that's great I do want to keep this brief. We got five minutes left. We got another question here So I want to make sure we address it. My name is Michael delzer for me The biggest issue is whether it be a service or distribution is the fact that I need people to be consistent on a common API in a common process Because I want to avoid the Enron effect because back in when Enron port clothes. No one thought the company would go down So I want to use multiple cloud vendors multiple cloud Solutions, but I need them to be identical I've tried using Amazon and it's a unique environment But it's very proprietary to them and it makes it difficult to have one kind of application delivery model to work in Amazon Because that's a totally unique ecosystem and the reason why I'm looking at open stack for American Airlines is I want the ability to have a Common application employment model independent of who it is actually running it for us cloud Foundry No, but I so I think this is this is the biggest challenge with open stack today And I think there's there's a lot of work that's going into ref stack into def court it to try to figure out interoperability between all the all the vendors From from my perspective I think we're almost thinking about the wrong way And if we look at something like platform as a service like cloud Foundry that can talk to Google that can talk to Amazon That can talk to open stack that can talk to VMware. You've now really solved that that multicloud API challenge I think there are still holes there in the service catalog like database caching queuing like those things aren't solved or consistent across little platforms But but that's how we we are thinking about it here for me the big that I'm planning on doing a cloud Foundry type pattern, but I also have to support the Microsoft people and Mongo is not going to be or not Mono is not going to be an operational viable for us So I still need someplace to host Microsoft type applications So I Echo one Jesse's comment I think when you have earlier when I said that you should focus on your applications I mean that you should focus on your applications so with you at least you're already looking at that and you're trying to figure out how you can get your apps to work across all the different clouds and In that sort of scenario paths is actually the best way to do it because you're letting the past be the one to do the integration you're just focusing on the apps, which is what your business actually requires and I know Jan you made a face when Cloud Foundry was mentioned I'm going to take this opportunity to say I'm still waiting for it Red Hat to come back and give me five OpenShift customers that I can go and talk to as part of my research We're gonna have to keep you two separate Look, we just got a few minutes left. There's another question here. Let's see. Let's get the question. We'll hit the homestretch here Just to follow up on that question between the service providers and the distro providers up there Whose responsibility is it to make sure that everyone who's Providing a cloud is providing the same set of API's the same set of services and adhering to ref stack and depth core and all of that Good question Anyone go ahead and tackle that one first. So I think there are a couple areas one is sort of the API side Where clearly OpenSec has the potential of becoming a defect or infrastructure for a lot of those Public cloud private cloud infrastructure today The reality is the major public cloud providers are not running on OpenStack So you don't get sort of the consistent API easily You could use OpenStack on top of say EC2 I would probably argue for a long time might but not be the ideal approach today. We believe please don't yeah Well, I tried to be nice But we believe there's a better layering on top of it The other part is also for example providing the capability to customers to deploy reliably their OS instances Between cloud providers get certification across cloud providers. So there's for example a program We have called the reted certified cloud provider programs are all cloud providers providing reted like in Amazon There's certified cloud provider and there's agreements around support SLA's etc So when you call up Amazon or call about somebody actually picks up the phone has guaranteed Response for example, and you can also move your subscription between these environments Seamlessly rather than having to I need to buy here. I need to buy here for example What you propose basically just locked you into red hat So now So now you're basically saying You're not proposing compatibility in OpenStack clouds. You just need to do red hat and no, that's not what I said That's what you said. You said just you choose a federation of red hat clouds and we'll get In all seriousness actually I think this is the foundation's responsibility Yeah, so and I am I am on the board so I am I can speak you know to do that So I think that actually, you know, there's been a lot of kind of red hat marketing probably here But I think that actually that's an accurate answer Because very the ref stack and the dev core and all of this foundation driven initiatives in my opinion to enforce API compatibility between different distributions and different vendors in my opinion is a lost cause and I think that Ultimately what's gonna happen is a certain number of vendors will evolve that roll their own distribution And with that distribution certain set of APIs, they just naturally comes in and This this these vendors was who's gonna be defining the standard not not not the foundation for like the ref stack initiative You know we got another question We're gonna tie it off after this one. I swear You just mentioned pass Closer speak louder, please. Yes interoperability problems. How do you convince SAP to run on cloud founding? Oh Yeah, I think they're actually a cloud found your foundation member. So they're looking at it How long will it take forever the man wants answers? He wants a solution. I'll write it tomorrow Yeah, I think that's a big challenge, but you look at And I call it cloud foundry in particular because it it to me mimics the the open stack foundation and its membership in a similar and government's model It's gonna take a while, but we're headed in the right direction in in that capability Let's just put it this way. I mean there's constant slugs, right? I mean what you do is basically there's this concept of an Unfair advantage from the startup world. So you basically sell a service because other people don't have the skills So obviously that will move things will get easier So your unfair advantage will come down and if you money that you ask for that is too much And people will move away and if you go to cloud foundry then there will be open shift and other people will go to open shift And then these are not interoperable and then it will fire up. I mean, it's just constant motion, right? I don't know just has got an unfair advantage because the rest of the market doesn't have the technical expertise, but Unfair advantage compared to whatever traditional enterprise sure I Say yes, so the I think that the operations model ultimately You know, you can charge a premium around the solutions that are new and hard like open stack And when it becomes mainstream and when it's down to three distributions, then you just this becomes the same business as managing the Immware cluster so it's a commodity business with a gazillion vendors in it that are you know fighting for thin margins So, you know, you can you can agree with that as somebody who's running that business But I mean I think the key for us is not to differentiate by adding people every time We had a customer the keys to differentiate and building a technology platform that gives us the ability to actually deliver this on a consistent basis As a product and to scale Revenue to scale installs without scaling headcount and that's exactly what we've done I think it's it's a fundamentally different approach than how anybody has has tried to do private cloud If you look at rack space private cloud if you look at IBM private cloud if you look at any that the other folks that are out there delivering a Hosted private cloud and the the implementation is very different than how we approach things Well, I think that the the this is like the key dish So where are you going? I think in my opinion is more like towards the Amazon model So you can automate everything can have a uniform data center with a uniform kind of orchestration layer on top of it And then and then and still is still you get commoditized to some extent But when it comes to managed offerings ultimately so some of the customer comes to you and they tell you I want a rack that looks like this I want three nodes with such and such configuration I want for you need people to go ahead and set up that environment You can have you know You can set up a data center full of uniform infrastructure and then automate everything out of it If that's what you do then you're going to the Amazon model not a managed services model All right, this is coming from the consulting company though All right guys, I think we're way over time now. I'm sorry Boris you can go punch him out in the hallway We'll take care of it right there All right, I want to thank our panelists for what a great discussion Lots of lots of food for thought and thank you for coming really appreciate it. Thank you Go grab some food and some drinks