 All right, so let's get started you guys want to move up a little bit. I Won't have to do a squint in the eye because of those lights. We don't buy it. I think we have can we get started Yeah, all right. Thank you everyone. Good morning. Hope you had a good time at the keynotes This is our first session of the day Along with some panel tracks, so I appreciate you guys coming here So the goal for today's session is to talk to our panelists about different delivery models that Are emerging because of the maturity of open stack and we can talk about some of the pros and cons of each We'll ask our panelists what their opinions are on how this Landscape is emerging So without further ado, let's get started So my name is Ashish Nadkarni I'm part of IDC's research team and my job is to look at market trends and track markets and You know size them up. So as as the tagline goes for IDC. We analyze the future A lot of people will tell you that open stack is one of the ways in which the future of IT is evolving So we can talk about what those trends are and how we can you know better Better embrace them So with that I would like to introduce our panelists. So I'm gonna allow you guys to introduce yourselves So good. Yes. Hi. My name is V. S. V as Joshi In the last almost like 12 months I've had two different and very distinct roles Initially for the first six months or so I was the founder and CEO of a startup and the startup was doing mobile and social apps And then in the last six months or so I have been with EMC and I have been with a Company that is providing solutions for those app developers solutions That will help them develop the modern apps and things like that So that is a dual kind of role that I have played in the last one year or so and I will forward to Cody Excellent. Yeah, my name is Cody Hill and I'm with platform nine my systems engineer and Very similar in the last six months multiple roles I was formerly the lead cloud architect at General Electric Supporting a pretty large private cloud there Thank you Jonathan Gerchater. I'm at Red Hat for nearly three years as you well know We have monetized OpenStack and open source across many products. I'm in the competitive marketing group there for OpenStack and Telco Excellent and thank you again as you know some others have joined since the introduction I Have the audience Mike So if you have any questions as we get into the panel raise your hand and I'll walk to you You know, I we only have 40 minutes, so I don't want to hold all the questions till the end You know as we go into the panel if you have any questions that you believe are important to get answered Let's jump right in so let me kind of turn to the panel. Yes. I want to start with you I'm intrigued by your dual role your app app dev perspective So tell us a little bit more about how you see this whole OpenStack ecosystem from an application Developers perspective, so it was a dual role, but a dual serial role the first role was of an application developer and You know as an application developer The things that we need are we need infrastructure that is available instantaneously. That is number one We need infrastructure that is going to scale up and down We need infrastructure that will allow us to start small and then grow big. We need the ability to Do frequent changes? So those are the kind of needs we need the application to be resilient so those are the kind of things we are expecting out of the infrastructure service provider, right and Think I mean again, I haven't talked about OpenStack at all as yet And these things can be provided by the public cloud or they can be provided by the private cloud or by the internal IT Departments as such within public cloud. Okay, we can go to AWS or we that we can go to rack space for open stack I mean for us we went with rack space not because they had open stack per se We went to rack space because they check box. I mean they checkmarked each and every of those needs that I mentioned earlier Yeah, so open stack is definitely one of the ways for people to reach there to get that kind of Infrastructure for their applications. Yeah, excellent. So yeah, go ahead Jonathan. Sure. Thanks I can echo your point because I think many enterprises when they look at open stack They need to realize that it's part of their overall cloud solution So you use open stack for your infrastructure. You need storage You need a cloud management solution you need a platform as a service for the application development and You can use cloud management as well to broker between Private and public cloud. So I think the main message is people need to realize that open stack is not the only part of your overall cloud solution Yeah, that's a good point. Cody. Did you have anything to add to that? No, I think they're both spot on Yeah, so a hybrid cloud strategy is very you know very good Even getting down to some businesses needing things like ironic in the open-stack product to provision bare metal and things like that. So You know bare metal virtualization on-prem Public cloud off-prem even a colo, right? You need to be looking at all these things when you look at a cloud solution And definitely at containers, right? Containers. Yeah. So I'm gonna do a bit of a role-play here So we has just hired his infrastructure guy Now the three of you are going to have to talk about your value proposition of you know Your own in your open-stack distributions or the pros and cons of what open-stack distributions They should be looking at what kind of infrastructure they should be looking at so Jonathan Why don't I start with you and you can talk about some of the? you know the different delivery models and What you believe are the pros and cons and then we can go in that order Okay, so you can certainly deploy open-stack in your own private cloud on your own hardware You can get it from a provider such as rack space where the red hat solution is available. You can get it from you know some companies offer a appliance model and With whichever the ones you're looking at what's needed I think most enterprises as we've seen this morning with the Gartner keynote is Open-stack is reaching enterprise adoption and The enterprise wants stability They want security and they want interoperability In open-stack was to bring this nirvana of removing vendor lock-in So they want to be able to interoperate with many hardware providers many neutron networking providers many storage providers Cody, what's the I'm sorry we Did you have anything? Yeah, so I think to add to that is that some of the things you need to look at when looking for An enterprise-grade open-stack solution is utilizing the hardware that you've already invested in Right, how do you turn that existing virtualization stack that you already have into a full-fledged private cloud with open-stack API's? And give you the choice of hypervisor, right? You want v-sphere. Do you want KVM? You're right. Let the IT shop choose the hypervisors that they want to use and and then you have to think further down the road It's not just hey, I want to be on Liberty today or metaka next week, right? It's how do I get there, right? How do you get from one version of open-stack to the other? How do we keep it highly available? How do you troubleshoot it? Patch it monitor maintain that distribution you need? You know there was a talk at one of the other summits about the six phd problem with open-stack, right? You need six phd's to run it and I think you know we can do better So so let me let me let me step back for a moment here. Yeah, let's go to the very basics as such in the sense that What what are the various? Open-stack deployment models. Yeah, so number one is okay fine I can go with public cloud offering like rack space and from them I can get open-stack. Yeah, and I go to open as opposed to Amazon I go to rack space and I go to open-stack because yes, it has the API API based services number one It has the the community support and these kind of summits and things like that And yes There are people who have gone through this thing before and you find these people all over the world And they have a lot of tools like the SDKs APIs and CLIs and all those kind of things So that is the reason why you come over here But now when you go to private cloud when you go to private cloud there are again three four options that open up One is you can build it yourself. You said, okay I go to open-stack.org and I can just download the software and I can build it myself That is one way of going about it. The second way of going about it is yes What Red Hat Mirantis and Canon equal does they provide a software for these folks and they essentially support that thing So that is the second way of going about it And the third way of going about it is open-stack in a box You can call it an appliance. You can call it as a rack. You can call it as an open-stack powered rack or whatever it is So and that open-stack Open-stack powered rack or open-stack powered appliance. So that is like Open-stack in a box where the vendors they have gone through the trouble of making all these hundreds of decisions That you as a do-it-yourself person has to go through and they have done those things They have certain view and they have put that views into that product and they are giving the product to you so that is the way I look at it and From the company standpoint, I think your question was how do you what do you talk about your solution? so from our company standpoint we We feel that the customer has spectrum of needs They can go either of these routes and whichever they're out they choose and they choose their route based upon what they're most Comfortable with if suppose somebody is most comfortable with do-it-yourself. Yes EMC has all the hardware gear and we have all the Drivers and things like that by which our hardware will work with all the things that you have if you want to go with The software model is we partner with Mirantis We partner with red hat and we partner with canonical and yes If in case you want to come up with an appliance, we are launching something next week not this week But next week at EMC world. So that is another route to go through so again I will not pick one versus the other I would say that we Meet the customer in their journey and whatever stage the customer is that yeah, and I'd like to add There's one deployment model that was kind of left out and that's the open stack as a service with your hardware on-prem a managed open stack And so that's and I know because that's that's what platform nine delivers right? So you have your open stack in the cloud and you bring your own hardware inside your private data center And you can wire that up and then all the maintenance and all of that's handled at the Service level, but isn't Cisco met up or also do the same thing Cisco metapod requires you to buy An appliance basically you have to buy their hardware and you need to allow the Them access to the data center to do the upgrades and the maintenance and all of that So this is kind of decoupling that right enterprises have tackled virtualization. They understand how to Upgrade vSphere upgrade their Linux hosts from KVM, right? And they but they don't know how to upgrade open stack and troubleshoot it and check out rabbit MQ and figure out all these nasty things Right, but virtualization enterprises have solved So that's that's where the line of demarcation is and then we host the open stack control plane for them and and that So I think that's the beauty of open stack is that you have this upstream project Where all these different vendors and delivery models should be adhering to and one of the things you want to look for whenever you're doing that is a Solution that is most adhering to that upstream so that Changes that are made upstream can flow down to the distribution or the version of open stack that you're using And then I think it's a very important factor to consider for especially companies that are no more leading edge Or want the newer features or if for example telcos are a big use case for network function virtualization And they'll want something like IPv6 pushed in upstream and then have that supported by a vendor So you want to look for a vendor that has also upstream Influence to get those changes pushed in and then support it So that's an excellent point and we'll get back to the whole community contributions But Cody I want to go back to you and kind of touch upon something you mentioned earlier and just now A lot of times we get questions as if So open stack is not always a green field green field deployment And I think what you're saying is that the way you deliver or offer open stack as a service You can almost make full use of your existing investments. Can you elaborate on that a little bit? Yeah, absolutely so with our product what you're able to do is For if you have a existing VMware deployment you deploy an appliance and not only do we wire up that those v-sphere clusters to be You know used as Nova compute drivers, but we also ingest all of your existing networks v-sphere templates virtual machines as open stack objects So you're not managing a legacy cloud and then moving forward And we do the same thing with KVM, right? If you have a KVM environment all of your bridges that you've been using with KVM come in as open stack networks All of the instances running in KVM come in any QCOW 2 images you have also come in as open stack images So we automatically ingest all of those And then allow you to just run and maintain one Distribution, so yeah, I completely agree because companies that have grown over years have many different vendors in their data center They have various Linux's they have various hardware vendors over time So just as you're saying if you really have an existing virtualization platform if you have a good cloud management solution that can manage your VMware and your Red Hat virtualization and your open stack and your public cloud Etc. You can embrace all of those technologies under one platform There are open hybrid cloud solutions such as when we offer a core cloud form that allows you to add an existing open stack on the side into your other solutions That's good to know and so going back to you. Yes in the context of the different delivery models in the choice you mentioned how does The use of existing resources your existing investments in infrastructure plain to the the choice that you are that you talked about Could you elaborate on that a little bit in the sense that it would seem like it's a similar Pattern to storage resources right or storage infrastructure You use storage virtualization to make most of existing storage resources But you buy a brand new storage array as a green field almost like a replacement to your existing resources Is there a similar model there in an appliance model? I think everything kind of comes together I mean the vendor has thought of all these things they have put their own kind of storage Devices compute devices and things like that. They have packaged it They're pre-tested it pre-validated and that's the thing that comes to your On your site and yes because they have done all these things You don't have to go through 500 decisions that you will have to go through So as far as that existing Existing infrastructure that you have yes, you can use it for the other models for the do it yourself model And for those things you can use because yes, you can get the drivers and things like that and you can make it More working with open stack as such but as far as the appliance is concerned. Yes You will get the whole thing packaged and pre-tested and pre-validated At your doorstep. So is there a way to have the appliance sit side by side and existing Kind of a DIY or a software based open stack deployment Do you believe that there is a way for both of these to sit side by side and for someone to make use of an orchestration layer at a higher level to You know use both site types of infrastructure. Yeah, I think yeah Do you guys want to elaborate on that or talk about that a little bit? We have quite a few customers that are using some of the hyper-converged stacks such as Nutanix and stuff like that So what if you're coming for the appliance model right Nutanix works great with VMware and vSphere Their KVM isn't full-fledged open KVM, right? So there are some issues in talking about a cropless. Yeah So, you know, we do see that and then you know, there's always ways to do, you know things through keystone integrations with multiple regions and stuff like that so you can keep your You know your appliance that you just bought your brand-new shiny appliance and then you know if you brought in something like You know roll your own or you know a managed open stack solution for the legacy You can always do like a keystone integration or something like that So you could have them running in multiple regions so you can manage them in the same location, right? Yeah, I just wanted to ask a question on the appliance model doesn't it if open stack is supposed to be The goal is I think you need to Hold your mic if open the goal one of the goals of open stack is to have more a vendor agnostic or Enabled you to use white box hardware Doesn't the appliance model bring back a little bit more lock-in like you locked into this box of Servers and every time you want to scale you have to buy a whole nother appliance versus just buying some few servers or racks of More standard x86. I think it's a it's a it's a good point. Yeah, there's a thing It depends upon the customer So if you have customers like CERN or AT&T Yahoo PayPal eBay for these customers They are not ever going to go with the appliance model because they want customization. They want tremendous customization They want customization to a nth degree. So they are not going to go with these things But then when you come to enterprises for the enterprise folks They don't have the they don't finding the talent that can do this thing I mean, you know open stack as such is a very complex thing There are 20 plus projects and there are 20 million plus lines of codes that is written over there Okay, the person has to go through hundreds of decisions as such Assembling the stack is a problem after you assemble the stack maintaining the stack is a problem the enterprise guy over there He wants the whole damn thing to run. That's what he wants He he has all the the VM related expertise as you mentioned earlier He has that thing what he doesn't have is somebody who can understand Python who has written something in Python It's what he doesn't have is so those are the things for which are The hardware expertise so for that they would say for an enterprise of certain size from let's say an enterprise Who have the capacity of half a rack to ten rack for those folks? Having somebody already figured out all these pain points already have a certain view and they are coming With with a solution that is the best way to go for them And that there's where I would kind of differ in the sense that yes There are some companies like CERN and AT&T Yahoo Yes, they will go for customization and they will need Open stack and they will not go for the appliance model But there are at the enterprise level Enterprise need these things to be hardened, you know enterprise need I mean they don't want somebody from their IT department To build the cloud and but then gee what happens if that guy leaves? Who's going to take care of this thing? Yeah vendors with supported solutions and professional services that are going to help you get you through your journey You know, you know the goal is not to go out and hire a hundred expensive Python engineers But get a solution that is well supported Very secure very stable That will be able to use white box hardware So this is a good lead-in to the my next set of questions, which is resources and I think we are on our 60 There's a question Michael Dells with American Airlines Well, one of the things you guys because I am an enterprise and I'm fortune 100 And when you say enterprise you kind of someone to offend me and because you sound like you all enterprises are identical But one of things I think you're not quite addressing is the business really doesn't care about us creating open stack The business wants to have a great way for them to deploy applications So for me one of the big challenges of a box solution where a distribution or doing it yourself is making sure that the vision is To include something like cloud foundry or docker integration or containerization type of Processes, but that also have the linkage to be socks compliant PCI compliant HIPAA compliant be able to think because there's Oracle pays their sales people really well So Oracle Exadata exologics all those kind of things that are really big boxes that maybe in the future will be open stack enabled And they kind of talk about it. The current product doesn't do that So that's what for me is the big thing is to figure out if I'm really small company And I can totally live in a box and I don't have to worry about legacy integration No big deal for the rest the enterprises though This is not going to be the entire ecosystem and I need to make that sure that ecosystem works with the rest My ecosystem which needs to work at a higher level not just that a VM type of Construct yeah, that was an excellent question. Thank you exactly my opening point is open stack is a sliver It's a part of your entire strategy and people shouldn't think of it as the be all and end all you have public cloud You have existing virtualization on on various platforms and you want to Entire solution including your past be it Open-shift or cloud foundry or whoever that should be open and be integrated, right? Yeah, and I think you made a great point You know I formerly I worked at General Electric lead cloud architects So understand the size of fortune 100 or even fortune 10 right as well as all of the compliance right with GE health care We had hippo, you know with GE capital we had socks and aviation and we have all these different regulatory requirements, so You know totally understand that model and there are you know the issue that we had with bringing in the appliance like and it wasn't your appliance it was At the time we were looking at Nutanix and different types of things just to make our lives easier and the issue is is that we have already Validated that we use this vendor for this type of hardware and if we change that we have to change all of our documentation Right, and we have to use this hypervisor vendor because we already made that decision Compliance signed off and we're done. We have to use that right So those are the things that we had to take into account when we started building an open-stack distributions We had used that hardware that hypervisor that storage and we needed a layer that can do all of that for us It's the ecosystem. Yeah, so I would like vS to respond and then I'm sure questions What you're looking for is multi services you want multi services from the same box Yeah What I would like to request you is you should come to the EMC booth and see the system that we have over there Okay, that is exactly what we are trying to go towards that. Yes. This is a box and yes to start with we will have the Open-stack based IS on it But yes, if you want pass you'll have pass if you want big data you'll have big data So the same system can be leveraged for various use cases. So it kind of Thanks for the question. I appreciate that. Yeah Hi, I'm Alex from America model Brazil. I would like to share Little bit of our experience down there in Brazil regarding not only appliances, but also Distributions, I would say that you know the the talk about appliances definitely we're heading to that But I I see I believe that we are a little bit Later when it comes to Distributions and a different harder mobiles just to mention one of the issues that we face it with a particular Distribution that didn't support hyperconverged infrastructure meaning we couldn't run VMs with our Storage and we and we needed to do that and if you look at the the marketplace You're gonna take canonical distro had had or Mirantis they Taking even the the same Release for example kilo the differences between what is supported in terms of Hardware and not not mention, you know Certification matrix I mentioned basic architectural decisions such as going with this separated storage or Hyperconverged it. We are still very E-mature in terms of what opens that from the vendors is supported or not To make a long story short we decided to go with the vanilla meaning we We are deploying ourselves and it requires lots of efforts The you know, there's no open-stack developers available all over the place But it proved to be the best way to go Because you don't have you know a terrogen. I mean a homogeneous Support for hardware and stuff Thank you, and I think so red hat has had a similar. I mean you've gone through this process, right? So, you know Yeah, so so one of the things that I just wanted to ask and he doesn't have a mic anymore But that's okay No, but I just wanted to know so with having a hard time hiring engineers and building a vanilla open-stack What is What's your plan to move off of kilo? you know and What are you doing for a maybe legacy? Legacy infrastructure that you already had virtualization on what are your plans for these things? I'm just I'm curious from a business point of view basically what we are doing and We do have both VMware. I mean we have a bunch of v-blocks. Can you imagine the v-blocks? Bunch of isolated VM or deployments We decided to go, you know from scratch. We're doing that and from kilo to Liberty or whatever We are evaluating to be honest with you our current distribution was upgraded from from had had to RDO and the developers I'm not a not a tag guy, but They're doing and we will support I mean we will be supported by developers and We joke. I mean I joke that our main developer. He needs to leave at our building He needs to ask for permission to leave the building go home and etc. So That's how we were working. It's a very risky situation and But honestly the experiences that we have with the distributions Was not a good thing and and we're running but you never know. Yeah, so That's exactly why platform 9 was founded. Thank you And so that's you know this situation where you have your hyper converged infrastructure He mentioned v-block right running on top of VMware And not having a team of people to be able to manage this right if his engineer gets hit by a bus Their whole business is down right. I hope that doesn't happen And so those are the things that I think a lot of enterprises, right? You guys are here to learn about how to deploy open stack and some of the things you need to look forward is how do we Maintain this how do we make sure that we can upgrade it? How can we patch it? How can we troubleshoot it and having one engineer that has to ask permission to go to the bathroom is You know, maybe not the best idea So we have a question here and just wanted to do a quick time check We have around 10 minutes, but this is your panel in the sense. Please feel free to ask questions I do want to talk about Resources and how you know, how many resources do you need to manage open stack and I think you're dust up on it, Cody Simon Alex on from America Mobile on the Hardware certified matrix what's important to remember is open stack is not running on the hardware there's something underneath it called a host Linux and We want to look for a Linux distro that has a wide variety of certified hardware and Has been a Linux distro that has been around for a long time with an ecosystem of Certified hardware and that that is a very important thing to remember when selecting that Open stack distro is what is the Linux that's running underneath it and additionally does that Linux have security? certifications, so for example the rail Linux has common criteria and Phipps and various other Certifications on the US government that's you know 20 years in existence We we can certainly have a discussion. Yes. I would say that would apply Yeah, we can debate certifications, but let's get to the question Hey, my name is Marcus from demand media. So we're in kind of the same boat as gentlemen over there We're trying to adopt open stack early on stages, but the way we decided to go about doing is more of a cautious more More cautious way. We're using metacloud at the point at current point But the early stage of metaclouds and not the whole metapod thing On our own gear But with the end goal of running open stack on our own the DIY type solution side-by-side with it If you guys had to give us just one piece of advice from each of you, what would you? See as the biggest recommendation or advice from going from using metapod Currently and then building a DIY on the side on our own Just trying to learn it and understand it at the same time I think the one advice I can give you is find a friend who has done this thing before Okay, I thank you. That's an excellent question. Yeah, I would highly recommend Making sure you have a deep talent pool to to help you do that, right? Start hiring and start hiring now if you want to run open stack on your own with vanilla You definitely need you know, it's doable No, there's a lot of companies doing it very successfully, but they have amazing engineers working for them So just start hiring Yeah, I think it leads into your question. You're gonna ask on resources and there are many cost benefits to Hiring on your own versus paying a vendor to support it for you And the resource constraint is really something you need to take into account and there are several Sorry to use the term graveyards of failed Deployments where companies try to go on their own and eventually had to come to vendors So I'm gonna throw this back at you guys and you all three have come from different background different suppliers How do you address the resources gap, you know today, you know that there is a big disparity between say what American Airlines would need to deploy versus the gentleman there Mobile yeah How do you address it? I mean is it education is a training program certification? You know documentation I'd love to hear your comments on how you would yeah, so The way that we address it is we'd use you know a managed open stack solution So we host the control plane decoupled from your hardware Enterprises today don't have an issue with virtualization like I mentioned before they know how to maintain and manage that And so we we upgrade the distribution we patch it we monitor it We troubleshoot it if if they try to provision an instance and it throws an error our engineers reach out to them and say Hey, what's going on looks like your storage is full right? So we have an army of engineers that are on our operations team that handles this for you as a managed service So that you really can just you know run your business do your thing and we handle that So again That is the reason why this whole appliance model exists right because yes the talent pool is not available And again, it's very simple guys. I think is you know, we are in the month of April We just did our taxes some of you did the taxes by themselves you use the DIY method Some people got the turbo tax software and did it by the again So you got the reference architecture from turbo tax. Yeah, and some people where you know I want to don't want to do anything you went to H&R and you said okay Do my taxes and they again so I would say that okay that is the way it is going to continue now You have to define as to what is it? What is your core competency is your core competency about? Installing this open stack and maintaining this open stack and yes There are some people who love to do that and yes go ahead. I mean that is what you will do And but there are some people for whom the velocity is the key velocity is the key and because velocity is the key they will build what they must and they will buy what they can and There's where the whole Appliance thing comes into picture. Yeah, I agree on core competency So obviously our model is that we curate and QA and test the code Extensively after we have made many contributions upstream So what you get is a solid product and with it professional services to help you deploy and to train your stuff All right, so Derek with solid far now net app So a question for all three of you can give me kind of three reasons a piece of why do you think some of the Appliance models for open stack have failed Specifically if you want to reference one someone like Nebula There's a good question. I don't know about the three part of that's that's nine Personally, I think Nebula was a fantastic thing, but it was just too early I think that's why Nebula failed, right? They came into the market too early They were trying to beat down the door and try to push something And and I think with an appliance model you much you have a much higher burn rate When it comes to a startup, I think that's why they failed if they would have come in as a software layer They might not have grown as quick they got popular people knew them But they just they burned through money and I think that was kind of their problem But no those guys at Nebula were smart guys. Yeah, I first handed knowledge on that so JPL is a customer of ours that we released this morning are running on our stack and they started with Nebula from Chris Kemp that came out of NASA and This the story that we are aware of is that they were taking open-stack bits from various distros so from upstream from some from ice health some from Havana and a Linux kernel from here and there and code from here and trying to get that all together to work proved very daunting and What they liked about our model is that we have the solid Linux and the curated Upstream open-stack co-engineered together and it's working great for them was reached today So I think as far as Nebula is concerned the Pedigree of the founders was absolutely not in question They had the best founder the person who kind of started this whole thing can be considered as the founding father of the whole Open-stack movement as such. Yeah I and again, I agree with Cody that yes They were a bit ahead of the time I guess in the sense that the surrounding market of the ecosystem had much not matured to a level where a Startup can sustain that kind of you know goes a startup They they are going out of cash and the map the market has to mature before they go out of cash And somehow that thing didn't happen in this particular case But because otherwise everything was fabulous, right? I mean that's that's all I can say I'm just adding more to what Cody mentioned Thank you We have one question and I'd like to give you time to do offer closing comments I'm Alka Dishpande from Oracle Corporation One thing you mentioned is to some some customers are doing this pulling Various features from various releases of Open-stack and that is a total Totally wrong. Wouldn't you say that? Cooling items, you know some features from various Various releases of Open-stack. I saw you know Yeah, yeah, when you're doing a this was an experience at NASA JPL when they were doing the DIY model is they're pulling directly from upstream to the the code has not been Extensively Q8 and tested with hardware and tested with Linux. So they ran into these problems. So any advice on How many releases are well supported or safe say I deliver something in Mitaka and then can I backport it to The previous release kilo and you know only two releases back ports are allowed. Is that right? So what you do mean what you're doing upstream on your own is really up to you If you're getting it from a vendor then they will have a support cycle So for example our support cycle is three years and other vendors will have their own So this is not a question about a particular vendor, but rather open-stack question I don't know if you guys are qualified to say that So I think the way I would sort of rephrase the question is I think she's talking about the Sort of the a la carte approach where you do have the ability to pull things push things With the Linux background, I think Jonathan maybe you can offer something about how Red Hat has managed that in a sort of Sort of the cadence and the discipline around it and you want to offer that and sure so yeah, so as People are well aware. We've monetized and run no open-source software for 20 years So we are very well versed in in running a rel and now you take open-stack And you run it on rel and you co-engineer the two together because open-stack relies on drivers to talk to the hardware to talk to the Networking to talk to the storage. So when you curate and you co-engineer the two together you come with a very solid and secure and Qa solution cool good good and I think to your point It's not always a bad thing to pull something from Mataka if it solves your problem But you need to QA it you need to test it before you roll it, right? So, you know In our organization, we've we've pulled stuff from You know, hey, we just had a bug filed and that bug was fixed in Mataka Well, we're not on Mataka yet, but hey, let's pull down that code. Let's put in our environment. Let's QA it Let's test it. Let's see what's going on So I think it's okay to do So I think the hope is and at least with with our platform is that we will then be upgrading to Mataka and that bug fix is then fixed So yes, you will be upgrading your distro But if you're trying to maintain Grizzly release with some Mataka patches, it's not a good idea. Yeah, so so with that we are almost out of time So I'd like to offer each of you the sort of you can your closing comments and you know What what words of advice would you offer to? users current and potential Offer. Thank you for the opportunity to be on the panel. I've enjoyed it and had the opportunity to answer the questions Certainly welcome to go on your own if that's your preferred approach if you want to use a hosted solution That's your should be open and willing to do that if you want to run it in your own Data center, then you want to really look for a solution that is solidly Q8 and tested across a wide variety of hardware and software platforms that Form an ecosystem. Thank you Yeah, I would I would mimic that same deal I think you know, we we have a lot of distributions running on red hat. We love that we have a red hat We love a boon to VMware if you're a VMware shop, right? if you need to choose what fits you the best and don't forget about legacy it because you don't want to build and This is an issue that I learned firsthand at GE is you don't want to build siloed teams to take care of different stacks of Infrastructure where you have a cloud team that's using virtual virtualization, then you have a legacy virtualization team So if you're gonna build your team to run your cloud moving forward combine the teams run it together and do that And and so and one of the things that we're doing to prove how open so open stacks gotten a lot of black eyes That it's it's difficult. It's hard. I can't I can't run it, right? so one of the things that we're doing to prove that open stack can be easy is that we We could run it for you, right? And you can download a free trial of our software and do all that or we put open stack on a stick Right, so stop by the booth get open stack on the stick and see how easy open stack can be All right. Yes. Thank you. So a couple of things number one is Velocity is the key second thing is your core competency if your core competency and If your core competency is about building it and yes fine go with that particular route, but if velocity is the key Yes, build whatever you must but buy whatever you can and Finally, you know when we say free software and things like that just remember one thing There's a difference between free beer and free puppy Free puppy you'll get a free puppy, but then you'll have to maintain you'll have to take care of the puppy And things like that. Yeah, so with that thought I think I should close. Thank you. Thank you I'd like to thank my our panelists here and thank you all for hanging in there and thank you for the great questions I have a great time at the summit