 I'm going to start. My name is Lou Tucker, I'm CTO of Cloud Computing at Cisco Systems. Every time I've been involved with the OpenStack community since the beginning, and one of the things is whenever I'm given an opportunity at these summits to give a presentation or whatever, to me it's an opportunity to more or less think about something different. In the last couple of days, right up to it to hastily put together a bunch of slides, which represent my latest sort of mowing around of different ideas. And so this is not going to be a technical talk. This is really much more going to be a talk about riffing sort of on a theme that has been going on for a while, which is this notion of a world of many clouds. We've seen cloud computing exploding across the IT industry and everything else. So one of the things I wanted to explore was as OpenStack gets more and more successful, what does the world look like when there's not just three or four public cloud providers running OpenStack, but there are hundreds. And hundreds and hundreds in different countries, and that what does that mean that when we have actually a world of OpenStack clouds, I'm glad to see Mark is here because also he was talking about cloud federation as well. So this is along the same kind of a topic of what happens when we have a lot of these clouds and is there anything we can learn from the internet? How we move from a network, how we move from a whole set of different networks to an internet. Anything we can learn from that as we move from sort of the isolated islands of clouds that we have today to maybe something called the inner cloud. So Juno is just out. Tenth release. I'll be really, really proud. Many of you I know we started getting back together in Austin or Cactus, Diablo was the first release I was involved in and through each one of these releases, one of the great things about it is that the number of contributors has increased, the number of companies contributing has increased and they've been getting better and better. By and large that's been true. I know that in some of these releases we did have problems and we continue to have a lot of issues that each new release I think improves. And so I wanted to first of all just applaud everybody and hope that we are spending the time that we get together because I think the reason why we've been able to go through this many releases in this short period of time is because we've got a very good structure for doing this. We've got a structure that every six months we get together in these design summits and we really try to hash these things out because in face to face we can move a lot more quickly and then on IRC and everything else we can go into a lot of the details. But so I urge those of you who are developers to really make sure you spend enough time in the design summits because we've got it, it's face time and we need to be able to make progress there. So Cisco has been a major contributor to OpenSack since the beginning as well and these are just a bunch of stats and the great thing about Stacklytics and I know everybody looks at it every day just like watching your stock price or something. But you can see how many contributors you have in your company and who's contributing into the different projects and since we started this in Cisco I've been watching that and we've been growing our team and we look at this a lot and I really encourage people we want to get up in the charts. The best way to also get up in the charts and this is where we need the help is in reviews and so what I'm really proud of is that we've been able to move the number of reviewers we have up in our organization because that means that they're reviewing other people's code. It's the best way to learn the code and a lot of times people have asked me well how do you get OpenSack developers you don't hire them you build them you grow them and so you have to have people in your own organization or get great guys out of college and put them on reviewing other people's code that way they learn OpenSack and they can provide a contribution very early in the process and that's worked very well for us in Cisco and we're working a lot of things not a big surprise we're working mostly in the neutron area and the networking area and this is in a lot of response to the change I think that's happening in networking. Networking below OpenSack is in a huge huge shift right now in SDN everybody's talking about overlays and underlays before everybody came in we're talking about IPv6 and whether IPv6 can be used we have a larger address space should we be using that larger address space to be doing a lot of the things around creating these these logical networks based upon segmenting that address space using that for route base you know flow base kind of isolation so which might be a really important concept for us going forward at the same time things such as colom and containers are important so we've had participation there as well. As a company we've been really in the business though providing solutions OpenSack solutions for our customers and this has been working with many of our partners being Canonical, SUSE and Red Hat bringing OpenSack solutions out into our customers obviously on Cisco infrastructure and that's where my team has been largely involved with that and you've seen a lot many of those customers up on stage during these presentations talking about their OpenSack implementations. What I'm most proud of is the fact that they are up on stage they're joining the community this is not a traditional vendor customer relationship where we ship hardware with software and that's it we want them involved in the community because that's what's going to be important to continue to move things forward and this is talking about one of our latest announcements that we did with Red Hat around an integrated UCS Red Hat OpenSack offering into the market. We also recently acquired Metaclabs. We're going to start to see a lot of this kind of consolidation we're seeing of these companies being acquired by others and Metaclabs was particularly interesting to us because it represents another sort of point or in this like go-to-market space this is about being able to deliver a OpenSack as a service so essentially the OpenSack is run on-premise remotely managed and by Cisco and so this makes it easier for people who don't want to have to learn how to do all the kind of management of OpenSack itself. We also recently announced InterCloud Fabric. This is helps us in terms of the hybrid model so that from an enterprise data center you can extend your cloud offering either onto Amazon or onto Azure or onto OpenSack and so it's a hybrid overlay model whereby from within an enterprise you can start to move workloads onto other people's clouds and there's also been a lot of news you've probably seen in the press everybody's now in a race of how many billion dollars you can invest in OpenSack and Cisco is not unlike the others in that and that we announced though and we are now moving into becoming a cloud provider but the tack that we're taking slightly different than others and that's what I wanted to talk to you about today is really we're looking as doing this with our partners in many ways Cisco wants to always work with our partners so we are providing now OpenSack technology at a number of different major service providers throughout the world and therefore we're bringing an OpenSack to what will turn out to be hundreds of data centers around the world whereby in whether they are in Europe or they're in Latin America or any place else they will be able to have a Cisco essentially run cloud through their own sales channels so they will be selling these services and on the back end they'll be run by Cisco running on OpenSack and there's been a lot of press and I'm sure you can look at that so in many ways cloud computing is turning into all of these different ways of delivering cloud we can deliver it in a public cloud we can deliver it as a private cloud or an on-premise cloud it's being delivered for IT applications Comcast is using OpenSack we've been working with them to deliver video applications Xfinity their X1 application so we're seeing it go into a variety of different different areas here I you know even in e-commerce that we're seeing Mercari Libra IOT management what I find interesting is that we're seeing a lot of management now moving to the cloud even though you're managing things back inside of an IT organization that's another area that I think is important so where do we see OpenSack going I know on the foundation and the technical committee we're having a lot of these discussions can we keep OpenSack as one thing while it covers a lot of these different use cases these different markets and I believe that we can and but it means that we have to focus on doing some things today and then keep our vision open for where this might go into the future and that's what I wanted to sort of explore today where it might go so one of the things that's very important is that we recognize you have to win the enterprise if we can't even though we may be saying this is all about the new cloud applications the cloud native applications IT shops really control most of the spend and so we have to be able to bring a cloud into these IT shops teach them how to move applications not necessarily the way they were but move what those applications were doing onto a cloud platform and because cloud computing is absolutely winning and it's the only real growth area right now in IT so within the foundation itself we have started these different working groups and this one is led by Intel and what we're really focusing on here we've got over 25 companies Cisco's own IT guys went to participate in this and then we're looking at things such as what do we need in terms of availability and how do we do monitoring of this how do we do you know charge backs the whole notion of cattle and pets I'm sure that we keep hearing I hate that analogy I don't know why I like cattle as well as like like pets but and then we and then we've also been kicking off these operators summits because very important these are people who are trying to use open stack there's a lot of us who are developing open-stack it open stack feels very different if you're a user of open-stack it's not so nice at times and so we're really trying to bring the operators together to get them closer to the whole community here consequently there's been a whole lot of blueprints and this is the process that we want this is a process I think that will can that will help us move forward you create a working group you gather use cases you you expose the problems the things that we need to do and you submit blueprints so that other people can start working on those things and this is how we work together as a community so there's a lot of blueprints now that are going in coming out of this application this win the enterprise area another area that we've seen is the NFV area this was a little surprising but in the last two years we've seen large segments of the whole telcon industry looking at the essentially the money that they're spending on single-purpose appliances in these core networking functions and so we've got a lot of actually and a lot of businesses are based upon shipping these appliances and are now looking to shift that to shifting to making these things deployable as virtual machines on top of open-stack so this is a completely sort of and one view you can say there's an orthogonal view of what you want open-stack to be but I see this is fitting in very nicely because we want to see these kinds of things merged into one platform to do this though not surprisingly what they're talking about they need to be able to dynamically provision resources well check open-stack does that we need to be able to do networking and we need to be able to do provisioning now we have to extend provisioning though outside of just the data center do provisioning also in the WAN as well we need to be able to have real-time response so latency becomes really important measuring things becomes really important we need high availability and as a matter of fact this is I find this funny one because we talk about this in sort of carrier grade open-stack well who wouldn't want carrier grade open-stack I know of every enterprise would also like to be just as reliable and available as anything else so this is right in line with what we're trying to do as well so I think that this is bringing up a lot of these areas that again we're forming several blueprints around and that there have been talks and panels I know in the show here talking about NFV but don't look at it just from the point of view of those of the networking guys who are worrying about how they're going to go and and build a net server or DNS or anything else think of these other things as well which are going to influence and hopefully improve what we're doing in open-stack I also think we're going to see a couple generations of these network services the simple ones now by and large or simply making a virtual machine form factor for particularly networking function running on an x86 box and everybody likes it says oh that'll drive down the cost of it and everything else that's not going to be sufficient the second generation these are going to come distributed systems and I think that's where it's going to be interesting so those of you particularly coming out of a background distributed system this is going to be an interesting place to watch because if you're going to have to support you know millions of net endpoints or whatever you're going to need a cluster of these things they're going to have to be able to share some state and they're going to have to be able to horizontally scale and that's not the way these boxes were designed originally but they have networking protocols that allow you to do that so it's I think it'll be a very fertile area of innovation in the next couple of years another one reaching back actually a couple years in this notion that well we've got all of these devices coming online so that we've been looking at it mostly from the infrastructure core focusing open-stack and how do we get this to work inside of a data center but we know that we're now connecting mobile devices handheld devices everything else and we know that there's even a larger swarm of these devices going out to the rest of the world it's interesting we've got some people here from Cisco have been working on looking at putting open-stack managing all of these different kinds of devices and and that I think is another interesting play ultimately you know we're also going to see it in these cars and again these are all network based systems there's so many devices now on a car with the networks and as these things get smarter and smarter where they're going to be they park themselves they're eventually going to be driving me to work and everything else this I want to be able to manage from someplace else like a cloud so in many areas I think what we're seeing is that more and more of these management systems are moving into the cloud many of you may know that we acquired a company Maraki which is management of Wi-Fi access points and that's now being done from a cloud and people love that I mean it's kind of it was puzzled to me at first why would would appear that way but it becomes an easier place to manage things that are spread out all over when you put them up into a cloud you know actually Apple made that shift a couple years ago I was surprised when they said that they no longer thought that the laptop or the PC would center the universe it's now iCloud because that's the only place that you can connect and synchronize all these different devices a new one which I think is really interesting and this actually doesn't have anything to with OpenStack except for the fact that Jesse Andrews is there and Troy Toman some of the original people inside of OpenSack are now working at a company called Planet Labs. Planet Labs is sending up a fleet of these little tiny satellites and essentially they're making a line scanner going around there in geosynchronous orbit so as the earth turns they're scanning it and in 24 hours they've got a complete picture of the earth as one you know and now what do they have to do with of course these are discrete pictures they have to be stitched together and that they can do things such as looking at water utilization and population growth at ground cover at forestry areas and then sell that information and I just happened to like the idea that it seems that now some of the early people inside of OpenSack are now working on this even though they're running this on Amazon but it's an eye but so I I like clouds you know I like cloud computing and that we are seeing that this is another example of we're going to enormous amount of data that we want to be sharing and they their ideas to democratize access to this picture of the earth for everybody on the planet and I hope they are able to fulfill that as well because half of that they want to make in the public domain so it's another area which I've really been interested in which is this notion of not just open systems but open data the more and more data we make freely available like this I think will mean there'll be more and more companies started up finding useful ways to package and sell and use that information so coming back to OpenSack really what is our future here I mean are we going to we're clearly on a trajectory to win in this marketplace and that we're seeing almost all the major IT vendors are now involved and so what happens when we have lots and lots and lots of these OpenSack clouds out there and should we start thinking about that so we can shape the blueprints and everything else we're going to for this kind of cloud to cloud interoperability so let's turn back the clock and look back at how this happened for the internet not sure how many in the audience remember the internet when it was this size I actually do is in 1977 okay a few of us or whatever and you could count the number of nodes or imps that were on this and where they were and you knew who the system admins were for these devices and we could send mail it was wonderful and we could share files it was wonderful and but we had no idea you know that that was going to become this fortunately a couple of major you know of sort of critical decisions were made early on that allowed that kind of growth and it wasn't completely as straight line as anybody's been involved in the node but we were able to get to this kind of scale so that now the rest of the world looked at this and said that's the internet you know it wasn't always that way before there used to be all these different networks and that you would have to be in one network and you moving from one network to another network there are different protocols involved and everything else like that and what we came up with instead through you know IP based systems TCP IP standards ITF work a lot of the work that's done by several people here in the audience to come up with these different protocols so that now even you know my 20-year-old daughter knows what HTTP is you know that's surprising at least she knows that's the thing you got to put in the front of something if you're going to you know get to Facebook or something so and the other attribute about this this wasn't done by one company here are the different companies like there's these autonomous systems that are connected so this is fierce competitors exchanging packets allowing other traffic flow to go through those networks so that you can create that internet and this is largely using protocol around BGP which means that you are way to exchange information so you can have essentially a mail system of routing packets throughout the internet so they go where they're supposed to go and this means that you've got AT&T working level three working with Verizon with Google trying to telecom Cisco and your companies as well through these through these different kinds of distributed computing protocols that allowed this to interoperate and send packets around to the world and therefore we've got you know huge huge networks of these things and this is you know very interesting graphs that can only be made and visualizations around it but also mistakes can happen and those of you who get CSCM a recent issue whatever was talking about some of these mistakes and why it's taken so long to get to secure internet routing and there's been some very interesting you know things at one point two-thirds of the internet YouTube traffic you know went through Pakistan and Pakistan was blocking it so went to a black hole no YouTube another point I think there are 18 minutes that China telecom hide essentially hijacked all of the traffic from Verizon both of these were believed to be human error mistakes that were made somebody put a wrong entry in terms of how much of the prefix that somebody should be looking at and all of a sudden the rest of the world's traffic ends up going into different places so it's not without problems but we've come a long way and it's still very much a work in progress but meanwhile we're able to use the internet so this gives me some hope even with OpenStack because I hear problems about OpenStack every day because we're deploying it and everything else I keep me that's okay it's okay it's gonna get better and we will always have problems but we have to just keep moving forward and because then look at how the internet's working today and even though we have these issues from time to time and it's a great article if you ever want to take a look at it so in clouds what do we have well today we have three major big elephants in cloud computing we got Google Amazon and Azure and you can see them in these kind of maps so the question is as we build up more and more OpenStack and we've got these other clouds is there a way for us to start thinking about this as becoming a cloud of clouds following that internet model these are different companies that is can be working together and what would it mean to have a cloud of clouds back at the early internet I don't think anybody I mean do you remember anybody even thinking what would you use ARPANET for I mean it wasn't wasn't really conceived of it being oh place that I'm going to be having Facebook friends and Twitter feeds and and everything else maybe Twitter actually was closer because I remember we did have talk and some other things that worked across it so OpenStack on a global scale and there's a there's a couple of good reasons why we want to do this first of all be nice to have it every place so that everywhere you went you can have an OpenStack cloud if we can work through all the processes that we're trying to do with dev core and around the the developing of what does it mean to be OpenStack you can have applications that can you a broader reach for application developers in terms of this and they can serve local markets because one of the things that is is we are still fiercely local people we are in France have you noticed it's different they speak a different language they have different privacy laws they have different notions around what kind of information they want to keep private so as we build these global clouds we have to respect those local conditions and so we want to sort of think globally and act locally we also want to be multi vendor one of the things I hear all the time as I'm talking to customers about OpenStack is that they don't want vendor lock-in as a vendor might make me feel bad or something but I respect that because I don't want a customer locked in for some arbitrary reason I want the customer getting a product for me because it's the best product I want to compete on the implementation and so that that whole notion can have a have an advantage here they should be going to the best cloud or the cloud that meets their requirements I would expect to see clouds that are OpenStack clouds catering to different markets catering to a high-performance computing market catering to a biology or pharma market where they have special real they've got special you know ability abilities to comply with different HIPAA rules around their data all of those kinds of things around OpenStack we can have lots of room for competition but having a common model means that you can have a larger market in which to to play on the developer side so we have to so it aligns very strongly I think with it with the essential notion of OpenStack but then we also have to sort of balance that because we want to have heterogeneity but we want to have heterogeneity is good because you have much more resilience if you're not all built on the same platform but you want to have the common platform so that you can have the largest application base targeting them so what will it take first of all it's a lot of business agreements and so Cisco is engaged in this right now with InterCloud and it's a lot of business negotiations and believe me the way the telephone systems work and in any of the internet work there is a lot of business relationships around peering behind it making it so that they can they can carry each other's traffic with cloud computing now it's not just networking but it's also compute it's storage it's customers identity management and things like that that we need to be shared we're going to need probably some intercloud protocols for that that's how that's how distributed systems work they have to be able to exchange information talk to each other in a common language to do these things between these clouds and I don't know what those protocols are and so I want us to start sort of thinking about them so that we can start designing them into the systems that would allow services marketplaces and exchanges to be built so that you could have companies that that develop applications put them into a app exchange or something out there and then consumed by any of these different clouds and we need then there are things such as Federation you know and policy also becomes really important here as well because right now a lot of the the way we enforce policy is through networking and through ACLs and things like that are very very specific to either a vendor solution or a particular data center so one of the things that the central question on this is that can we do this you know can we have this kind of intercloud based on this kind of community driven model instead of waiting for one of those large service providers to come out with this global cloud and I would want to put my bets on the community as being able to do this and and embrace them and bring them into that community as well so with Cisco that's what we are we're beginning to with intercloud we're working with first and foremost looking at the different applications particularly the Cisco already has and SAS applications around Webex scan safe and a lot of other applications putting them on this cloud working with our partners and working with even linkages into the public clouds so that you can start to have this is a very very early prototype let's say of what we're thinking about it and I've had it described as an alliance between Cisco and partners so that we can make this happen so some of the work that we've been doing directly in open stack around this has been first of all in terms of federated identity that becomes a keystone you have to be able to start there you have to be able to have keystone SAML assertions and everything else so that you can have users logging into different clouds single sign on is just so old school you have to have that at this point and then can we start to move up the network again because when we get around to policy you much want to capture developer intent this application should only be accessible from either these points in the world you know I don't want this accessible from this other geography I want it only accessed by these people I want it only using these networks and that can only be done by having a sort of a meta language and so there's a lot of work like in Congress and another thing trying to define open stack policy in open stack we've got in group policy API is that are being discussed in the implementation of that that should be able to work on this as well and so there's some other links here first and foremost we've got to get open stack adopted and therefore we have to get these issues around stability we have to be able to become trusted and trusted both in terms of availability but also in terms of security I really don't want to see vulnerabilities you know exposed on open stack that would kill this notion very very quickly so it's important that we do that and then I think we have to start think about networking in the broader sense stop looking at it as a data center problem it really has to be a multi data center issue we have to look at the WAN and we have to look at the emerging technologies in networking today to make this effective so in summary sort of like you know could the internet have been built by one company you know I really don't think so it's not likely one company could have done it there were many that were trying to and it was only when we really developed the technologies that allowed multiple companies to cooperate that we got to the growth of the internet and so we'll open stack be the community that builds its inner cloud or will it be a single company or some of the other large providers today and I think that's really up to us and so that's why I want to sort of bring up this topic now so that we can start to think about that as we go forward and designing the next set of features for open stack as well so that's it thank you very much and I will take any questions we might have I think I've got 10 minutes wonderful great or any other debates we want to have we'll start with questions about this topic anybody else been thinking about inner cloud or cloud of clouds or federation yes yeah maybe come I think that's an important question could you repeat it then for or let me rephrase it I think the question was how do I see the impact of politics on any this notion of inner cloud obviously it's kind of a great influence and I think that we're seeing the struggles even now on the internet particularly around I know Google's having problems in certain countries so politics is going to always have a play here and I think I don't have an easy answer for that I don't think we're going to change the world and becoming everybody is going to be the same so we have to accommodate local differences and so simple case around data sovereignty you know yes guess what if you're in Europe you're probably going to want your data and not to leave Europe so you might want you might be want computation or the execution to go someplace else but your your promise to your customers is that their data is going to stay in country and it's very and if we build that in early on I think we have a chance of being able to enforce that kind of policy politics is I think going to be very important so you give analogy interesting analogy of internet and intercloud so obvious one so in case of internet you had a governing body like IETF which had a substantial role in determining the success of internet so two part questions what kind of governing body do you imagine for an intercloud because you would need something like that otherwise things won't automatically happen and then the second part question is you have companies like AWS Google Azure you mentioned which have to play a very integral role in participating in these kind of governing bodies so how do you view this kind of body be able to corral these kind of companies so that everybody goes and works on a common vision of intercloud that's a great question I may want to have Fred comment on it as well because he's been a part of IETF one of the the only minor comment I make about it is that is actually it was less around governance it was around actually request for comments so maybe Fred can you give us any of the responses that Fred Baker has been involved for those of you that don't know me I was the chair of the IETF for five years and been involved in this just a bit I don't know you want to pay me to do that it wasn't about governance it wasn't like anyone sat down and said thou shalt do it this way or well actually there was somebody that sat down and said thou shalt do it it was the operators of various and sundry now workers said you know I've got IPX I've got Apple talk I've got no IP do IP and basically pushed all their vendors in that direction as far as governance we had two structures one of them was the internet assigned number authority which you've probably been hearing about in the press for the last year trying to figure out who hands out IP addresses who handle hands out top level domains and that kind of thing you know so it's a working through those issues and they set up what now called the NRO the regional internet registries that hand out addresses they very much took a service perspective on my members our companies I give them resources in terms of address space and they make the business work and the companies do whatever it is that they do I'm not telling them what to do I'm enabling them to do it with the ITF we were developing protocols and once again enabling communications okay not for some purpose or you know thou shalt do it in my way but enabling people to innovate and to move ahead and then what happened after that was that all the different companies came in and said okay I'm going to build on this base I'm going to innovate in my way this needs to come back to the community this I I can take as my own and we're able to move ahead and the internet as we see it today came out of that process a whole lot of elbows and a whole lot of working together what I'd really like to see an open stack is exactly the same that's great and I think we have part of those processes in place today I mean that's why we've got blueprints which is a proposal that we use it to rally people who want to work on that I think what's different is that we're approaching this and maybe and this is sort of the shift I think in the sort of standards bodies work versus open source we're building in the software and I think that what we're trying what we have to make what we have to recognize for to get interoperability you have to understand you still need the standards and you still need the interoperability and still need the APIs there has never been a problem between open source and standards okay open source is a way to develop things standards is about what to develop and if we can work together it actually works pretty well five minutes left yes so for identity federation you mentioned you can't really do this without it you know there's there's various and sundry identity stores and multiple providers networks and and you're talking about keystone and what it means to open stock can you sort of elaborate a bit there and how they might hook together simple answer no but it's a problem that we need to address and I don't think it stops just with the use of SAML or any assertions around who I am that you you have to say okay which authority like you'd mentioned so we have to build support many of them we just want keystone as a framework to be plugged into that just like we're plugged into Active Directory or anything else we should be able to do that minimal I also worked at sales force for a while and did app exchange there and one of the things that you need to also recognize is that you often need delegation and to operate as if you are somebody else because if you're providing a third-party service that is providing services into let's say a VM that's owned by a particular person you have to be able to have privileges on that so we really need a much fuller our back kind of approach in keystone where you can have chains of delegated permissions on this and keep it simple at the same time so I think there's a so that's why I was pointing out that work that's being done because I had urged people to really spend the time to let's try to make keystone work I think that that's one of the issues that we're having if it's not to fit if it's not solving problems the right problems for different service providers then they're going to do their own and I think having a common identity system is really important particularly if you want to get up to this level as well other questions yeah go ahead so what are the key points to handle inter-cloud traffic so what point is discussed in working group I think there are many types of data flow in current services so what do you think so if the question is what kind of of do we need a common mechanism to pass traffic I think we already have that and that's why we've got TCP IP fortunately the or are you able to talk to another service on the internet uh we have that today so we can build on that framework and so we'd be arrested the very common way for you to be able to use somebody else's service without knowing anything about how it's implemented inside maybe your question there's getting more to the fact that when we start talking about very big flows if you're moving we know that that I love the analogy of saying that the highest bandwidth network we know today is FedEx delivery you know because they're shipping disks and the network is lagging behind the in the growth of data is happening faster than the the availability of bandwidth and that's creating well sorts of problems around data gravity data once it's created in a place tends to stay in a place data rest and stay at rest because it's too hard to move if we have a better notion of a integrated graded cloud of clouds I think we can start to exploit more and more linkages between that and increase that and solve some of these problems but to me that that's totally a new area for are there alternative means of transporting large amounts of data and then you've got on the other side well what about really really tiny amounts of data it seems like in networking we seem to oscillate between trying to solve one problem or the or the other and we need much more kind of unified approach probably one more question yeah what would you say should be the base set of capabilities for an inter cloud if we were design assistant today where would we start in terms of what does it need to do that's really a hard problem hard question so what are the base capabilities I think it starts with how do you allow use again that's why it's what it's identity federated identity being able to have a customer go from one be enabled provision if you've got a relationship between two different cloud providers and they say okay a customer of cloud a should be able to use services on cloud b with the same identity that would be the that's table stakes you got to be able to do that the second is that you have to be able to do service advertisement I think what are the are you running the same set of open sac services there as I am running here so that you can create profiles of your applications to know whether your application is going to be able to be run there some of that is intersects with the dev core stuff that we're trying to do to define those core capabilities so if we've got an ability to have a user in another cloud you know what the cloud is capable of then I think it gets down to it has to make economic sense so if to do this without recognizing these providers need to make money and therefore they need to do billing cross billing you know so there's an economic basis that we need to start worrying about and those might be might be already in place you know we have them for other mechanism we have them for the banking system we have them for the airline system we have them for a lot of other places where different companies work together and maybe we just need to apply them and adapt them here but so if it's users capabilities a way to make money or bill bill for things I think we could we could start there okay that's that's it thank you very much appreciate you your attendance