 Hi everyone, thank you for joining us for the panelist session with the panel session with the analyst My name is Caroline McCrory. I'll be moderating this session Yes, there you go. I am quite short. I can't you can there you go now. Not all of you can see me, right? so I used to be had a product at Piston cloud and then I went off to giga ohm which most of you know giga ohm shut down Very sad. I'm sure a lot of you miss it from the RSS feeds. So I'm actually very glad to be back with my fellow news and analyst folks to discuss a few things about open stack So I would like you to start by introducing yourselves Hi, I'm Lauren Lachal. I work for over an industry nice company based in the UK Which is part of the informal group, which is a huge conglomerate about two billion dollar Insights and I special I've been industrialist for about 25 years in the past four years I've been specializing in cloud My name is Frédéric Lardinois. I write for TechCrunch a Silicon Valley base block that's now owned by Verizon I don't know how that happened, but I write about the cloud developer tools and Tendencies more on the technical side than than anything else Good afternoon. I'm Al Sadoski. I work for four or five one research I'm responsible for the service provider channel. So everybody that does private public cloud hosting of all sorts and then I am also focused on this small little thing called open stack and pretty much All the different business models and market size Related to that This was a docker panel. Am I in the wrong room? No I'm Shawn Michael Kerr. I'm a senior editor at e-week. I hope which is a Quinn Street Enterprise Publication also right for internet news, datamation, server watch Linux planet a few others. I also manage a site called Linux today core focus for me is Linux cloud application development and Security stuff and happy to be here So the first question that I have is quite tame one and I'll start I'll start gently so It's open stack good for startups and you know innovation or is it just for the big players and major sponsors of the foundation and I think I'll pick on Frederick for this one because you write about startups I guess there's two aspects to that. Is it good for startups in the ecosystem and I would say Yes, absolutely. I mean it seems to be doing you know going pretty well right now There's plenty of startups doing all right, but there's also a risk I think with a lot of the big players now entering the market and you know the IBMs and everybody else joining joining the game as well So there's a risk there the other Aspect of it is I think is is it a good platform for startups as well to use and It seems like there's a lot to be done there still I don't know of a lot of startups that are betting on open stack as a platform for their applications It's definitely more on the enterprise side at this point I think but I see more startups trying to join the ecosystem rather than startups were using open stack for their business Absolutely, I mean are you seeing the same thing? Yeah, I was just going to ask you to clarify the question But you just did and thank you But as far as a startup finding a niche within the open stack community to offer a product or service There's obviously plenty of opportunity there It's a you know growing market as we see over the at these events every six months The attendance gets bigger and the show the marketplace area gets larger. So there's clearly some momentum So plenty of opportunities. There's companies that built their business around one specific open stack project You know, there's many many have done that. I'm not going to call out anybody specifically in fear of forgetting someone But there's clearly opportunity, but the other part of the question about is some of the bigger vendors Sure, they have a lot of influence One large service provider has nine Ptls So clearly they have a lot of influence on the direction of of some of the project. So yes, I would agree that Well, I like what you you bring up a good point with the with the Ptls So how do you feel about some of the the Ptls who also like for example, Do you think the foundation is doing a good job with trying to get all the voices heard not just from the community But also from the people who are in the ecosystem trying to build a business Ptls are really just cat herders So in an open source community, they don't actually mandate anything They just kind of line everything up and then it's always should always be a meritocracy So it's for the people that have been Ptls. Usually they tell me they do it because they have to I think the open stack foundation is gonna really good job of transparency of making sure that there's open levels of Contribution and then it's fair across multiple things but more importantly what we see even even this morning I talked about, you know, the app catalog one of the annoying things for me is app catalog you could get Marano Apps you could get a heat template or you get glance images All three of the same kind of sort of maybe so if there is a project that somebody doesn't like because of Ptl Or whatever they just go start their own So there's going to be a lot of concurrent parallel efforts and that's neither good nor bad. It's just choice Well, I Concur with my fellow panelists then yes I think the foundation is is because of the meritocracy that the pins the whole process It's not whether or not just for working for a startup is whether or not as an individual who contribute to the whole development in terms of Startup using open stack differently At the moment startups don't get any money to for their own internal IT they start on the public cloud so open stack would be a platform not for a pure startup but from a start for a startup which has Gone through some growth perhaps needs its own infrastructure for whatever reason And so we'll then develop its own Cloud which then can be integrated with with public cloud I mean Zynga is a good example of a of of a long-standing a company which started small on the Amazon Web Services then developed some private cloud and now kind of has a workload lifecycle approach to To the way it uses the cloud when a game starts it starts on Amazon Web Services if it's successful It's moved to the to their Private cloud I think the private cloud is cloud stack But it could very well be open stack and then towards the end of the of the the life of the game It goes back to to die a graceful death on on on Amazon Web Services. So I think that's that's how I I see Your question, okay So Sean you raised a really good point about PTL's and you know meritocracy and the projects and you know Governance ish and thereabouts. So do you think that a tighter core of open stack projects, you know Like a ref stack, you know, like for example with compute network and storage Something that needs to be really stable before other projects can be around can wrap around it like an ecosystem Do you think that's a good idea? Is that something that open stack should focus on it is that what they're doing, right? That's kind of sort of what Defcore is but the confusion is Even though the initial bunch of projects or you know compute Storage networking maybe and dashboard. There's other bits and those aren't core So yeah, they're doing the right thing by focusing on core core bits and that's fine When I think about it from a Linux perspective though, Linux is a little bit different Linux You have user space user space is user space once it's in user space. It's always maintained and it goes I don't know if Defcore will work that way or not if it follows the Linux model. That would be great If it doesn't not so great, but this is only the first release coming up. So we'll see Al what do you think of it? The only thing I would add I think it creates some complexity for the distro providers to market their particular version So they have the you know regular six-month cadence of releases for the core now You have these other satellite projects that may have a different release cycle and Now they have to figure out When they are going to cut it off to form a package that they're now going to support so instead of having You know vendor Release based on kilo now. They got to think of maybe going to a release cycle That's less than or greater than six months so that they can Line up some of those satellite projects into a into a supported That many enterprises would that would be upgrading something like that every six months and enterprises that we talked to That's one of the knocks on on open stack is that the it's too frequent The release cycle so that they like to you know hold off maybe a year or even longer They're it's what they're used to with the legacy software that they've used right Patrick you were to me it feels like it's Assigned that the project is maturing in many ways as well that that's becoming an issue at this point But it's up. I mean open stack is so complex at this point that you do need Some focus that just to keep it under in check basically because there's so many site projects and so many little things happening That you focus on the three core areas. I think makes perfect sense Well, that's this two elements to deaf core the first element is marketing It's about slapping open stack logo onto your distribution And I think that's that's good for everybody in the sense that you know what you get And from a vendor point of view also, it's it's kind of part of the marketing arsenal From a technology point of view. It's about stability reliability, etc so I think again, that's that's a win and I would have like deaf core to evolve a bit Faster because the first the first sophistication. It's it's early next next year But I think they've caused really evolve nicely because last year it was very much Trying to boil down the ocean very philosophical and in the mean in the neck last few months It's become much more pragmatic. It's about, you know, let's get on with it start with with the three core resources And and and do it so I think it's at the way it's been It's evolved from management review was improved for marketing point of view. Yeah, why not? It's it's a good development and from a technical point of view is also. I think it's a good development So it's a win-win at all levels. Okay, so my next question is really around And I what you want I want I'm interested in Sean's perspective on this one. So putting on the spot there What is OpenStack's biggest barrier to adoption in mid-market and enterprise now we've touched on a few things The biggest barrier in mid-market is obviously complexity and I think when we're talking mid-market It's in many cases. It's probably more than what a lot of organizations need Which is probably not what people want to hear in the bubble here at OpenStack But it reminds me a lot of the early Java days because early days and Java Java was be all an end all Yeah, J2E everybody plugged in and then at some point it got less complex with Tomcat But then you know people start using no JavaScript PHP other things other kind of compiled languages I think the same thing will happen here So there's a lot of layers of complexity management and organizational pieces in OpenStack That are not necessarily required for smaller scale organizations I think if you have that's the beauty of docker and containers you have a host you split it up You don't need all those other Abstraction layers that OpenStack provides so I think OpenStack in the mid and small-sized business market as a as a consumer Maybe if you're gonna roll your own forget it. It's way too much complexity for those signs of organizations So and you're talking about it from a complexity of using it as your own infrastructure as a service and deployment Yeah, because I can run I can get a to-you server run it in my office or in a data center run Docker Scale it up and put a thousand virtual machines on it. I'm not gonna do that with OpenStack There's too much overhead. I think the upgrades we already touched on the upgrade cycles is probably one of the big things I mean we've heard from Xerox and they're using Folsom and they said that they don't ever want to upgrade I don't know if they've changed it now It's been a year since they came out into the public and said they've got a thousand plus nodes running on Folsom. So What do you think? Yes for the same question, but do you think application portability is an issue for people who already have legacy apps I want to port it to OpenStack, but there's no real way of doing it So I've and I've had this discussion with some of the vendors today that in the case of VMware, you know, you can deploy your application in vCloud Air you can move it to a hosting provider or Cloud provider that's Based on VMware and you can pretty much drag and drop it works fine. There is fear uncertainty and doubt that you can take a OpenStack-based workload or an application and move it from your private cloud to a OpenStack-based public cloud or a private cloud with no interruption and in I think a lot of those things like federated identity Help but there's still some work to do to ensure that an enterprise is not going to have an issue and not be locked In to any particular one option I thought Lauren, do you look like you had something you might want to you might want to say on that one Yeah, no, I just thinking federated entities is not quite I would see that as part of that problem But it's true that that complexity is an issue and what surprised me in the past few months has been the explosion in Managed OpenStack. So so the first stage in managing the complexity has been the distribution Which is there to to to to make the the upstream code much easier to ingest But even the distribution have been difficult to ingest because of the complexity as well as the majority of OpenStack And now you You have all sorts of managed Offering from platform 9 which is completely SAS to meet a cloud which is on-premise, but completely standardized to I don't know Miran, sorry Yeah, I don't know these guys but To Miran says It's more of a custom one So yeah, so from that point of view. I think that's going to to to make to make Adoption a little bit easier from from that point of view my only point because the second question here was on application portability from the Diablo release on there was there's east there was an EC2 layer So compatibility even in the earliest days you could always capture an image as a glance image Then take it and put it on another workload. So application portability is kind of built-in always It's more what you choose to deploy as your own infrastructure Which you want to manage which is you know talking about the management is more complex and everybody's gonna say their management They're gonna make it easier. No one's gonna say they're gonna make it harder But but yeah, it's application portability. I have to take is as a given unless you're yeah Just take it as a given like you said earlier right this glance There's this that's that like which which is the right way to do it, you know, yeah So so I think I think I think we might we might be almost so be saying that So that an open stack projects need to get out of each other's way so that they can make open stack as adoption a success Wouldn't you say what would you say would be a good measure of open stack adoption success? What would sort of criteria would you for Drake put on it? It's a number of applications that are ported on it a number of nodes What is your criteria for measuring success number of production environments? I think at this point we're seeing more and more of them But I think just a question of who is adopting it and how you know how many people are adopting it And especially maybe in the mid-market, you know coming going down from the enterprise level just a question who's Complexity and everything we just talked about that. It's gonna be hard You know that I don't see it happening in the next year or two Well from your perspective for our audience, would you say like if you have at least two apps running on in Production on open stack that's considered successful. Would you say 25 or are you gonna go based on number of notes like what Xerox does? What's a good way for them to try to quantify there? You know, what would be a good success criteria that we could give them for for the open stack community for If they were looking at open stack, you know, they're to them. What would you say is good adoption? A couple of fortune 500 companies running production environments, you know, maybe 10 maybe a dozen companies moving over I'll jump in so number one just increased like every open stack summit has exceeded attendance records each one can so Consistently doing that obviously shows some momentum But one of the things that 451 research has is the open stack market monitor We we actually have bottoms up estimates from over 70 vendors across a bunch of different business models So if those numbers Revenue wise from vendors who are here sponsoring this event They want to make money if that revenue continues to go up. I think that is probably the Very good measure of success whether it's from a distro provider Or a Hosted private cloud provider a public cloud provider or a training vendor, you know There's a number of different business models within the open stack ecosystem if they continue to grow bigger than the general market Then I think that's a probably the best measure of success for open stack. Yeah, I mean for from a market point of view Yes, the success of open stack is a given. I mean, let's face it when you have HP IBM or a call Cisco All the telco and all the telco providers you are way Ericsson Nokia Alcatel Lucent everybody's betting the house on open stack. It is going to happen So that that's that's you know fine in terms of add success from I would say an internal Consumption point of view. I would say it's when opens there's two element First is when open stack is moving out of its IT niche to to be the platform that the line of business people are going to to go to naturally and I think the application catalog Launched this this morning will be an it's a nice development because basically it enables IT to go to the line of business people and Say, okay here. This is what you can run on open stack So it kind of accelerates the the adoption the second kind of niche that open stack needs to get out of so first first Niche it's IT. The second niche is workload at the moment. Open stack is mostly used for Kind of website type thing and mostly and bit of development As Open stack Becomes more important. It needs it's going to run much more enterprise workload OSS BSS on the telco side Business application on the enterprise side So it's when when it's Expanding from this IT niche to to be something that the the business audience will will turn to naturally and from a Workload niche to become a much more generic Operating system for the data center Then I think that that'll be a good reflection of its success I just want to jump in because I know again. We're in the bubble. So I'll take the devil's advocate position There's a lot of people here, which is great. You'll still see triple that at AWS reinvent, right? Yeah So there's still more AWS recently just broke out their numbers five billion dollars What's the number for open stack you would know better than I would but it's not five billion It's it's it's not so is success for open stack being as big as Amazon I don't know. That's a question. Maybe we've talked about but what's interesting to me is the first open stack vendor I'll call it rack space is now no longer betting the farm on open stack In their recent third-quarter analysts called they were saying how they're trying to de-emphasize that and focus just on support across multiple forms of infrastructure including potentially Amazon Microsoft 360 etc So I think that's very interesting to see that a vendor that bought bet the farm five years ago on the verge of Bankruptcy may get it be getting bought out. Who knows is now They own the farm exactly are now Hedging their bets and I think that's a leading indicator for what may be to come I think you know adoption participation is fantastic, but the money side. I think Rackspace may be the canary in the coal mine here Well, the most of the rack space farm is still running on open stack So the fact that they are Diversifying is more of kind of business model decision To go after several markets rather than a decision to kind of move away from open stack the same way nibbula failure to As as an enterprise is this is the failure of nibbula not the failure of open stack. So I don't So I don't quite buy that that I want to change the discussion How many of you in here are actual enterprise users of open stack? Hands up just me show hands because there's a there's a reason for my question, right? So so There if we look at come enterprises where they have a revenue of 25 million dollars and up they run on average 500 applications the reason why I asked what would you guys as industry experts feel would be a measure of success for an enterprise Who's using open stack was it one app in production on open stack or two or three when if the average is 500 You know, what is good? What is a good measurement of success for anyone who's trying to justify running open stack as an enterprise? We understand as vendors who are trying to do it and get into the need into the niche and into the business and everything else But what what does it really mean for somebody who is in an enterprise who has a P&L who needs to run? Services is open stack really something that they can but their farm on and And naturally do that is it mature enough for them to be able to to trust not just one app But more of their apps to actually run on it and Frederick There's plenty of telcos and everybody else who are running a lot of their apps on open stack already at this point, so If you want a number, I don't know 25 percent or something like that if their apps on open stack I'm not sure I'm not sure it matters all that much. Well, I think it's more of a question of how many How many customers are they supporting through open stack that yeah for telcos? I agree. That's a good that's a good measure You know success criteria, but I'm talking more for like it from an enterprise perspective Oh, what do you see a lot of enterprises? What would you recommend to them for 42 for if they were to adopt it? 42 is the answer to the answer to my question We can go home hitchhiker's guide to the Galatia reference I don't think there's an exact answer for any one enterprise. There's some enterprises that have made the decision that by 2012 a hundred percent of their apps are going to be running an open stack. They've made that Requirement, there's others that are just now kicking the tires and they have dev ops Only in open stack and are running all their production in VMware like there's no single right answer for enterprise It really is I'll take a question So the question was for the sake of the recording would open stack be a de facto standard for private cloud And will they also end up being a threat and would it be largely irrelevant in public cloud? I? Think some would argue that it is the de facto standard for open-source private cloud now But the mware still has a pretty sizable market share currently, but I Don't I don't know if I would say that it is going to be the de facto for all private But definitely the de facto for any open-source private clouds So my I think AWS is the de facto standard for public at this point Yeah I would just agree with that on the on the private side if we take this big tent model where everybody is welcome including Microsoft and VMware then everybody already is Open stack if Citrix Microsoft VM where are all Supported in a private open stack kind of distribution for API's then it's it's done it everything is open stack. It's the big tent Everybody's welcome. It's just how you divide up pieces Good so the comment from the floor was that open-stack needs to make it a lot more Stable also has good performance metrics and also a way for them to actually measure true performance just like the Google Release that's allowed to do that Yes So Right, so this is why this is this is but this is this is this raises the point So the the the points that was given by my IBM friend over there was that PayPal user does users open stack for their payment processing and it's basically one app It is that does that can constitute not being a success all being a success and I suspect I'm going to get a Counter here from the front of the audience So so the counter was if it took them twice as much infrastructure as far as hardware network and people To actually do their payments processing on open stack versus if they didn't then it may not have been a success So I think we're clear that success criteria varies based on your wallet There was there was a we had an industry analyst session afternoon yesterday where the Foundation got us to talk with quite a few Companies one of them. I don't remember what they tend to you're not I'm not going to mention their name But differently what they had done because they were they are a financial institution is that they are they had Costed the actual cost base Prior to open stack down according to the guy we talked to to the cost of the cable running on the on the floor and and then afterwards with with open stack and it was you know more than half the end result was Much lower cost with open stack than it was with a Prior architecture. So it's it's it's indeed depends on how well you architect open stack. It also depends on how well you understand your cost base And some people will will be more Able to kind of understand that than others. I'm not sure but So the answer was that there was operational cost included in the But it's a good question. Yes, you need to do that as well. So there was a question on that side. Yes The larger markets or the small the mid-market The talco market, okay So the question was when would open stack embrace the talco market, which is a you know billions of billions of dollars as embraced the talco market Lock stock and barrel. I mean you are way as in as I embraced So as I was saying you are way Erikson Alcatel Lucent Nokia who else I mean all the big telco That's that let me clarify So so what you're saying is is when it opens that going to understand better Telco operating models and factor that into to the take off and embrace that. Okay, just Very quickly You're in terms of global issue. You're right. We're talking about open stack being the Linux of of Infrastructure as a service It's not the big difference between the open stack and Linux is that enterprises don't contribute to Linux and But the like of people and Comcast contribute huge amount to open stack. Do they have a voice? in the foundation which Truly reflects their their their contribution answer No, and that's one of the the failing of the I mean I think the foundation is doing a good job at that level It's one of the area They should do much better and I've been telling that that telling them that for a while So it's it's not just telco. It's a more generic thing, but it's your right from a telco as well it's kind of compounded by the fact that there is a IT versus telco universe convergence and and not yet the mentality of not yet kind of Adapted to that convergence situation. Okay. Wait, so hold on. So you have something to add and then I'm going to come to you Exactly Thank you, okay, so it's good to see you again by the way, okay, and then wait so you so So So just to repeat the question most a lot of companies are going with private cloud because of the security element and they Is it fud or is it real or would they regret the decision and I'll give that to our because he's been dying to say something Analysts not wanting to talk So think back a number of years With co-location People wanted to deploy their servers somewhere near where their office was, you know the server huggers and as they evolve to cloud That's kind of there, you know next step. They want they don't want to give up and Put it out into a hosted environment. They still want, you know a little bit of access Compliance regulatory is definitely a big piece of that and Based on our like voice of the enterprise research, which now covers over 10,000 users 30% of cloud budgets are Private cloud Focused so there is a lot of interest Backed up by data that shows that people want to build private clouds and that is their foundation for a hybrid environment We like to say the road to hybrid is private and There's a foundation of private Coupled with you know another private cloud or public cloud to build a hybrid environment and showing them Because it's security that's come up. Yeah, the security is hard whether it's public or private and it really it all comes down to configuration so sometimes It's whether it's private or public so long as it's managed because no individual organization has the skills resources and talent to do Security properly today, which is another point because we only have five minutes left I think security is the largest single Underdiscussed topic in OpenStack and it is a ridiculously huge issue And I think it is obscene that there isn't a key management project as part of deaf court Which means all these connections between the different pieces you're connecting from horizon to Keystone to this and that over SSL TLS Who's managing the keys? I don't know that's not defined. I don't care whether it's private or public That's the question that needs to be answered with anyone that's deploying So my closing question which you all have one bullet each that you can fire What is the most? Annoying thing that a vendor will not talk about they're unwilling to talk about when you ask them That because that's a good nugget because analysts always get one question that analyst will never answer. So I'll start with Lauren One just one Go before we end That's the one thing you that annoys you the most that a vendor will not answer the one question that you have that they won't answer Because they usually answer my questions Okay, Frederick Real usage numbers and user numbers. There you go Al I was gonna say the same thing like how many customers how much revenue by specific Exactly the metrics that have dollar signs next to them typically Failures nobody wants to tell me about failures. Everyone wants to tell me about lighthouse wins I want to know when things failed and then how they fix them and rarely do I get that? Well, thank you very much for for coming and it was nice to get like questions from the floor. Thank you