 I think we're ready to get going, although I think we may still be a few minutes early. But it also appears to be standing room only, so there's a chair over here if you want. Come on. Okay, so to make sure you're all in the right place, this is Enterprise Vendors in the OpenStack ecosystem. Okay, cool. Really, I've got some seed questions, but we want to make this as interactive as possible. There's microphones throughout the room, or if you shout really loudly I can repeat your question. On stage with us we have in no particular order, particularly because they're not actually sitting in the order that I've put them up there. You just want to kind of go from the microphone at the far end all the way this way. Introduce yourself, who you're with, and then we'll get going. I'm Mark McLaughlin from EMC. I'm the other Mark McLaughlin. I believe there's a red hat number one contributor to the foundation that's not me. But I'm the EMC guy. My name is Jim Walker. I'm Director of Product Marketing at Hortonworks. If you're not familiar with Hortonworks, we are one of a few Apache Hadoop distributions that will now run on top of OpenStack. Robert Esker with NetApp, been leading our efforts, development efforts the last couple years. By the way, thanks for the clarification. I explained a lot when I was asking the Oslo questions to you earlier. I guess you're not the Oslo PTL. Apologies. I'm Andre Pash with Arista Networks, and I run our OpenStack development efforts there. And I'm Pete Yamasaki with AMD. I'm Director of Product Management, and we're part of the Server Solutions BU, which is C-Micro, which was acquired by AMD one year ago. Good afternoon. My name is Delia Storup. I'm work brocade in a product management role. I've been looking at OpenStack since pretty much its inception a couple of years ago, and we are basically building solutions that are best optimized for OpenStack at brocade. Right, and we'll talk more about that in just a minute. Okay, so we've got a pretty good selection of folks that are in the ecosystem or involved in the ecosystem. Do you want me to start with seed questions, or do you guys have anything? Come on. You're all here for something. All right. So let's start with where do you guys see the ecosystem in, say, the next five years or so? So, like, let me rephrase. How do you see the current state of the OpenStack ecosystem, and where do you see that evolving to in five years? Okay, so I guess I'm going to get started. All right. This time will work closest for me back the other way. All right. So basically, I think that, well, it's really amazing to see where we are at today, right? Take a look at this OpenStack initiative. It started, you know, back in 2010 sometime, and just a couple of years later, we are here with OpenStack, at this OpenStack summit, gathering 3,000 people, and also with a huge, rich, and very broad ecosystem that is already in place, right? There is lots of companies out there that are building business models around OpenStack. Some of them looking at OpenStack from purely a software perspective. Some companies are building hardware-based solutions, but also you see companies basically taking a stand in terms of support services and generating services, and obviously also some system integrator basically embracing OpenStack and offering their services. So I would say I'm amazed. Just a couple of years later, we are here with this rich, broad ecosystem. However, I would say that most likely, you know, we are just at the beginning of the journey. There is a lot that remains to be done, and in my view, there is lots that we need to focus on as a team in terms of bringing OpenStack in the enterprise, making OpenStack much more consumable by the enterprise segment, right? And what I'm getting at is this OpenStack is still very much this infrastructure as a service orchestration engine, right? Basically a set of APIs that are low-level that talks to the underlying infrastructure. OpenStack, in my view, needs to move up the stack of the cloud orchestration framework and getting to a place where we can have also performance monitoring, as well as resource quota monitoring for VM instances. Wouldn't it be great to have a system that would tell you that, let's say in a few days, all your VMs are going to be exhausting the capacity in the infrastructure, building accounting, service catalog, and service portals. All those components basically need to be created to bring OpenStack in the enterprise and to make it much more consumable. And that's where I see this ecosystem helping out. About the next five years or so. Yeah. Okay. Pretty much. Pete. Thanks. Well, I haven't been involved as long as you have. This is my third OpenStack summit event, but I've seen a significant change. So where would I say the state of the industry is today? Well, the one I went to two years, a year and a half ago in San Francisco, what I really felt was most of the people attending were developers. There were a few people actually looking at real deployments. But what I saw really change was a summit after that down in San Diego, where I started to see a lot of people who were really interested in actually trying to build infrastructure as a service based on OpenStack. And now we come to Portland and there's even more people. So where do I see this being in five years? Where do I hope to see it? Well, simply I'd like to see OpenStack as the de facto open cloud operating system. And for that being the case, there's some predictions I'd like to make here. So in five years, where do we want to be? Number one, I see it a lot like Linux. Where Linux is today. Where Linux is the open operating system. There's a number of supported distributions out there. It's easy to install, free to use initially. But there's also companies out there who provide enterprise level support and distributions for those who really want to run a supported application based on OpenStack. Next thing I want to see is seamless federation. And I think this is what OpenStack allows, is the ability to seamlessly federate between your private and public clouds. So this is one of the reasons we've been working with RackSpace is I think they start bringing that a little closer to the truth, is now they're actually certifying the hardware. They're building a distribution that they use in-house and providing as private clouds. That's going to help with that federation. Thirdly, I think we're going to see a lot of innovation from both software as well as hardware providers. This has been a little bit harder to do with some of the closed cloud solutions because it's harder for the hardware innovators to bring new innovations. They can't develop, they can't submit their own changes to OpenStack. And lastly, I think we're going to see a real thriving community of these hardware vendors and software vendors. And as he said before, we'd like to see it move up the stack, go beyond infrastructure as a service where people are now deploying not just a piece of software that runs on a server, it's an entire application that you can deploy as software as a service or platform on a service on top of OpenStack. I'll go quick here. I'm just kind of echoing your sediment there. I think what's been really exciting looking just at what our customers have been asking of us over the past couple years and looking forward is that it's clear that OpenStack has really arrived in our customer demands where they're not just looking at it and experimenting with it, but also looking to really deploy it. And this has evolved in terms of the type of customer that's willing to make the investment in deploying it from very large cloud providers to now even smaller enterprises. And I think part of that is reflective of the cooperation and partnerships among people to put together OpenStack solutions like Rackspaces certification and others. And I think looking forward, I won't make too many predictions, but I think that's certainly something I see continuing to make it easy to deploy and manage and automatically provision and grow your OpenStack clouds. So the... I believe once upon a time a guy named Bill Gates said that nobody would need more than 640K memory ever. So I'm going to go ahead and avoid the five-year prediction. But as a general role, the evolution of OpenStack seems to require, in order to be accepted and consumable, repeatable, become more of a known quantity and allow mere mortals to actually deploy, as needed at this evolutionary stage. It's needed something like a Rackspace cloud private software. And frankly, there are other distributions as well. I think we'll play a relevant role here. So in moving to a point where there are these known quantity distributions, where I think entering a new phase where we're going to see much broader adoption of OpenStack, I think it's also important to kind of think of OpenStack, not just as the whole entity, but maybe also the abstractions, the APIs that it presents. So if, for example, Cinder becomes kind of the industry standard mechanism, the provisioning control plane for accessing block storage, it may not necessarily be OpenStack itself. It's something that evolved from OpenStack. And as such, any development you're putting in upstream is going to be more broadly applicable than OpenStack itself. So I think that's an important consideration. I think just an observation around, I guess, rationalization of movement to cloud. There'd been a big rush, and here's cloud washing became a term. Let's hit that check box. I have something. It was sort of shared. It was sort of multi-tenant. That's my cloud. Or maybe, hey, Amazon, I heard the thing. I read the thing in CIO Magazine, and I got to have my AWS, whatever that means. Well, there's a next step, which is, hey, what does this actually afford me? And I think most folks kind of see hyperscale cloud providers for the value of instantaneous availability. I get a credit card, start my development project now, built it against elastic infrastructure. I don't really know what my workload will amount to, so this is a perfect place to put it. But the rationalization exercise is, over time, as you understand that workload, the economics may favor moving it back on-prem. And so what is the mechanism for doing that? And OpenStack seems to be the obvious reason, the obvious way to do that. Imagine kind of a broad deployment of OpenStack-based clouds around the world in the true promise of hybrid, the ability to burst back and forth between them, I think OpenStack avails that. Five years is an entirely way too long time to predict, especially in this space. OpenStack, you know, I represent a company that's in the Apache Hadoop world, where Hadoop was a year ago to where it is today, is a completely different world. And I think what's the same thing is happening in the OpenStack space. I think if you look at where we were a year ago to where we are today, it's completely changed. And so for us, we feel that, you know, Hadoop is kind of a killer app for OpenStack. I think they're kind of marching hand-in-hand together over a cliff, and really kind of is the future of a lot of what's going on, a lot of the enterprises. Hadoop is representing net-new workloads in organizations, and where is the best place to put these net-new workloads if you have scale-out, compute, network, and storage? It's obviously virtualizing in the cloud. And if you look at one OpenStack piece with another OpenStack piece, it all makes sense and goes together. I think the biggest challenges to both Hadoop and to OpenStack are what I call the 10 or 12 questions that are in every RFP that I've ever answered in my entire life, redundancy, reliability, security, all that core stuff that we all have to have. And I think those are the things that we're going to have to get past in order for everything to be adopted. I see kind of OpenStack kind of where EMC is, so to speak, in terms of we're just getting started. In terms of EMC's perspective with OpenStack as we've contributed to Grizzly, some kind of baseline functionality. But over the next six months as we move towards Havana, you're going to see a lot more EMC activity, just like I think you're going to see a lot more, of course, OpenStack activity. And I agree 100% with the comments from that app because of the accelerant to the OpenStack fire will be through the distros, through programs like Rackspace and others, so as large do-it-yourself customers are relatively self-sustaining in terms of skills to be able to deploy something like this. But as the distros start to do the certification and the support, I think it's going to accelerate significantly into the enterprise space, more so into Havana, and that's also coincides with how EMC is going to considerably ramp up our participation as well. Okay. So there's a microphone right there if you don't mind. Just a question for each one of you. Can you give an example of one feature that the company has given or contributed to OpenStack and one feature or something, some kind of product that you guys built of OpenStack that you pay for, like your customers would benefit from having your company. Does it have a difference between what do you guys do for the community and what do you guys do as a value added? Great. So I'm on the spot. Thanks a lot. Tag, you're in. It's easier to go first. Just like I said, EMC is starting off just like OpenStack and I would say we're 10% or 20% into it. So to be perfectly honest, we're starting right now with basically what we call platforms and protocols. So from a roadmap perspective, the way we manage our roadmap is we're going to fill out our platforms and we're going to fill it out across the protocols. Okay. So one of the things that we contributed to the community to answer your first question was a presentation that was given earlier this morning by someone from our team, Edgar St. Pierre, who is involved with Fiber Channel. All right. So we're working with Brocade and let's see, IBM and HP, of course, on contributing Fiber Channel. So that's contributing into the community. In terms of value added capabilities from EMC, you don't see that quite yet. You're going to see that in Havana. And in Havana, you're going to start to see greater exploitation of kind of data services from EMC around HA and DR. Okay. So I'm going to leave it kind of at that point. But there's a number of announcements. Some of them are coming out at EMC World, which I hope to see most of you out next month. But there'll be some very important announcements at EMC World that will have an OpenStack vein running through it. For us, there's two sides of that question. What have we contributed? There's a lot of work we're doing in terms of creating an HDFS, so HDFS, a highly distributed file system for Hadoop, distributed storage, and HDFS to Swift Connector. A lot of work going on in that area right now. The other area is really around deployment options and how do you deploy Hadoop and simplify deployment of Hadoop on OpenStack. So we're doing a lot of work within the community on that side. What do we sell? We don't sell software. We sell support and services. So every single thing we write goes back into an open-source project, be it in the Apache Software Foundation or within OpenStack. Every line of code we write goes back into the community. That's part of our DNA. So NAPS involvement in terms of contribution submissions started in the Essex release. And we debuted what at the time were referred to as NOAA volume drivers that since evolved into a separate service referred to as Cinder. That's grown over time. So we initially enabled one mode of operation and now we have several different deployment choices. And in particular, we have a kind of a next generation sort of mode of operation for our storage controllers called clustered on tap that we're exploiting. The basic theme is to avail core NetApp capabilities within the context of OpenStack, make them accessible through the abstraction that Cinder presents. So that's not necessarily something that we're not selling that. We're enabling the usage of NetApp with OpenStack to empower OpenStack to access the capabilities that are otherwise valuable to folks when they elect to deploy NetApp. In terms of other contributions, we actually just this week are announcing that we've submitted a prototype to add an entirely new service to OpenStack. And there's still some community discussion around the form that that ought to take, but the code is available as a work in process or progress submission in Garrett. Basically it implements file shares as a service. So provides access to shared file systems amongst tenants in much the same way that Cinder is a provisioning control plane for block storage. This allows you to request and receive coordinated access to shared file systems across tenants. Or in the more modular sense, you need not necessarily be tenants, but other consumers may be more for non-virtualized systems. Programmatically get access to existing shared file systems or create net new ones. And it's built to be sufficiently abstract to apply to NFS, SIFS, varieties thereof, perhaps an HDFS, more of a distributed file system capability, Luster. The list goes on. That's a net new strategic addition to OpenStack and not a basic enablement, if you will. We have also a tremendous amount of interest in what SWIFT represents. It's somewhat of a philosophical question. You know, is SWIFT only a service or is SWIFT an API? It's our position that SWIFT presents a de facto standard API and you could take a look at our portfolio and see that there are object storage capabilities and perhaps in the future, those dots will connect. So, that's a bit about where we're at. So, like EMC, I think at RISC our contributions are really just starting now and ramping up. Our main focus has been on how to integrate physical network orchestration and monitoring into quantum. So, quantum, as you know, is really good at orchestrating the virtual network, but we really want to go a step further and be able to take the information in the virtual network, apply it to the physical network and really truly get end-to-end provisioning of both your virtual and physical network in your cloud deployment. So, that's where our contributions are really focused on right now and to plug our design summit it's tomorrow afternoon on that topic. In terms of where our value is, you know, I think that we're trying to, we kind of have two value propositions. One is really just being the best physical network infrastructure in our cloud on, but really where we provider value within OpenStack is working with partners who provide OpenStack solutions and providing a platform and unique capabilities with which they can provide a better OpenStack solution. So, as an example, we have multiple OpenStack providers who are actually running their software directly on top of EOS, our operating system on the top-of-rack switch. Better auto provisioning and auto configuration as part of that OpenStack cloud. And so I think that's really where the kind of value that, you know, customers pay for is, if that makes sense. Thanks, so I think we're going to be in a little bit, my company, AMDC Micro, we're in a slightly different position than the other vendors because what we've really done, and let me start with, I think the value that we bring to OpenStack is to a degree we've reinvented the server. First is we, you know, change the game in terms of providing better space and power. And the second thing is we've converged a lot of infrastructure elements into a single fabric platform. We converge servers, networking and storage. So for us, it's this capability and how do we provide, how do we take this value and, you know, bring it back into the community. So the first thing we're focusing on, you know, we don't have, you know, we're still a small company, we're a startup that was recently acquired and we're in the process of growing. So we're taking baby steps right now. And the first thing we want to do is how do we take this converged architecture and how do we make it consumable? How do we make it useful for OpenStack? So the first thing we're concentrating on is now that you have all these elements, you can almost have a cloud in the box or you can tie it to external systems. So we have shared storage within our system. So how can we make the shared storage available? Some people want targets through EMC or NetApp, through Cinder outside, but in some cases they're going to want to deploy the whole thing in the box. So what we've been working on is how do we include things like the Cinder API into our management layer? And what we're not trying to do is we're not trying to change it. We want to provide a standards-based API for configuring our storage so people can use something they're familiar with, something their tools are familiar with rather than integrating to one of our systems. So that's very powerful. So what we're doing right now is trying to make it easy to deploy all of OpenStack on our system. The second step is now how do we take this infrastructure and make it useful? In our system, our today's system, we have a 10RU system that has 256 atom servers or 64 Xeon servers in 10RU. So if we follow the OpenStack model, you can deploy virtual machines using Nova Compute. Now what if we took Nova Compute, and now we could take selective hardware nodes and use the same mechanisms that you use to deploy VMs and also deploy physical machines? And we haven't contributed anything yet into the community. We've been working on this and we want to work with the community to deploy something that's going to be industry standard. But for us, we see it as because of the architecture we're bringing, there's some unique value add that we want to bring in phase two. So I'd like to highlight that Brocade, Brocade communication system, joined the Alliance back in May 2011. So that's more than two years ago, right? So as from there on, we're already as a team, as a company, we've already started to clearly identify what it means for us, for networking company to be part of this Alliance. And obviously what we've been doing, like some of our partners here from the networking side, we've been starting to basically consume OpenStack, right? It's the easy piece to get down. There is Quantum API that a company can take and basically leverage to create interoperability with your networking platform and solution. So we just contributed back our Brocade virtual cluster switch plugin, which enables Ethernet Fabric services to be completely orchestrated by OpenStack. So, and again, you know, it's a contribution that is very much focused on consuming OpenStack and the work that this community has such as been doing. We've got some extension that we've done in that context, a little bit on the line of what Arista mentioned with a different approach, right? On the IP side, we are also consuming OpenStack to create interoperability with the LogBalanceService API that just got folded under Quantum. So we've got our own product line for load balancing, and we are actually the first in this community to basically have a prototype running for orchestrating Layer 4 to Layer 7 services, again, for OpenStack Quantum. Now, the piece that is more interesting is really the work that you are doing with EMC. I think it's, you know, Brocade has been championing this effort and we started to work on that and we want to go already. What we've done is basically taking a hard look at how we can enable the CINDOR component to be extended in such a way that not only you can actually manage iSCSI but you can also manage fiber channel storage and fiber channel sand. So six months ago, we had this idea. We reached out to a tier one storage vendor, EMC, IBM and a few others and we started to do a weekly call to basically come up with the design of what it means to have fiber channel support in OpenStack. So I would say that the results are really speaking for themselves. In the grocery release, the fiber channel storage work has been done. So those extensions are there with multiple plugins provided by vendors, right? And in the Havana release, Brocade is going to be taking the lead in terms of implementing those abstraction and representation of fiber channel sand services into OpenStack. It basically boils down to a fabric, an FCSAN zone manager, right? Plus obviously we're going to add our own plugin as well as part of the Havana release. So I think we are definitely embracing OpenStack at Brocade and we are also very much willing to come with the community and have the community to really get to a stage where OpenStack is really, you know, I would say really has this broad coverage in terms of technology on the storage side but on the IP side as well. Okay. Oh, come on. It's a packed room with standing room only. There's got to be some questions, guys. Really, we're going to turn this into like the interpretive dance session here if you're not careful. As I turn my own microphone off. So I did have a couple other canned questions. The one I was most interested in is customer use cases, how things, how you're handling hybrid or bursty workloads or what store you have to tell in that area. If you don't have one now, what is your direction looking forward? And then let's start at the far end and come this way again. Okay. So we have probably eight or ten interesting use cases of different variation that we are working with customers on. So a lot of times what happens is as you, as accounts have questions, fundamentally kind of starts with what's EMC's strategy and approach to OpenStack. And so we help a lot through those kinds of use cases. A lot of use cases are basically, can this investment that I've been making in EMC infrastructure over the last ten years, you know, can I use that as a part of my private cloud deployment? And the answer is yes. And so we spend a lot of time on that use case, which is basically maintaining kind of legacy interoperability with your install base. So we talk a lot about that. And then kind of moving forward, we're doing a lot of work around with distribution partners such as, you know, such as Rackspace as well in terms of kind of moving forward, net new deployments of private clouds or hybrid cloud models. Can EMC part of our, be a part of our strategic strategy there? So customers come to us all the time saying I'm interested in deploying private cloud, moving my applications, you know, hosting them in a cloud environment, whatever. Can you be a part of that? The answer is yes, of course. So a lot of the use cases really have to do with kind of maintaining your legacy investment and then being comfortable that you strategically can continue to think about OpenStack and EMC. And that's kind of really kind of blocking intact on what we're doing. So Hadoop is an application on top of OpenStack. I won't bore you with the data world. But I mean people are looking to use OpenStack because there are really three different reasons within the Hadoop world. First of all, it's provisioning services very quickly. Being able to spin a cluster up, spin a cluster down very quickly. You know, like for instance our engineering team will work on a 70 node cluster during the week to do development. You know, five days a week that things split across five different teams. And then on the weekend we're running performance testing on that thing. That's the most easiest way I could think about provisioning and reusing infrastructure for various different purposes. But there's a lot of reasons why organizations would want to spin up a cluster or spin a cluster down to do some sort of data science in some short term, right? It really comes down to the temporal temporal, man I can't talk today. The temporal nature of some jobs that are being run in Hadoop and being able to easily provision a cluster? First day of the conference and you're already having like your nose up at 4 a.m. to catch a flight. That's the first reason why people are turning to OpenStack to run Hadoop. Really it's the provisioning side. The second side is, well it's the elastic nature, right? So being able to add compute and storage to Hadoop cluster, well Hadoop is compute and storage, right? So it's really kind of part and parcel to where organizations are going. So as they expand and as they get more data and as they ingest more data they want to scale rapidly. They want to be able to scale rapidly without actually having to redeploy a cluster or make massive modifications to the Hadoop cluster, right? So the Hadoop cluster typically has X amount of nodes. We don't really want to care about the bare metal underneath, right? They want to separate those two sides, right? It eases deployment. And then the third part really is the whole multi-tenant nature of which, you know, Hadoop has its own issues, right? And then Hadoop, right now within the Hadoop infrastructure it's kind of difficult to actually, you know, segment data off within a particular cluster. Well if you're able to virtualize, you're able to spin up a cluster, separate it out in a virtualized environment, set up your firewalls around that you're going to be much better off from a security point of view. It allows us to perform charge backs. It allows us to provision out clusters to part of the organization, have them pay, right? And these are the kind of things that are going on in the Hadoop world. Hadoop is also very new to organizations. They're struggling to get their hands around it. And if you think about it from a deployment point of view, it's already a challenge because, well, we're starting off with something that's complex and new to the organization. You're trying to figure out here how does it fit into your data center. We just really want to make that much more easier and you start to see why the two of them kind of go along together because of, well, quite honestly the workloads that are kind of being developed around the organization. So I hope that helps. So the use case is when I've been thinking about the last as these two other gentlemen spoke to eloquently, it's a little difficult to describe because we see interest and adoption deployment across nearly every vertical and nearly every geography in very different use cases. I mean, perhaps it's an oversimplication to state that if you have infrastructure and you put it into an as a service model, OpenStack facilitates that. And so, granted, there is a lot of, I guess you could probably say there's been some predominant trends, certainly analytics, clouds, you know, big data. You know, I mentioned the desire to facilitate a, apologies for the term, AWS style development workflow but do it on-prem for reasons of IP or economics or several other reasons. Actually, national sovereignty for example in Europe. So, apologies for the cop out but it seems to be everywhere. However, on the the notion of hybrid, apologies, so on the notion of hybrid actually we've doing quite a lot of work around the concept in general and it's certainly applicable to OpenStack. So, you know, within our suite of capabilities, our portfolio, we have something called SNAP mirror which is a thin replication technology and it's particularly interesting when you think about like hybrid use cases, you know, you think about I guess cloud as a utility and usually it's kind of characterizes an electrical utility. You turn the switch and it's on but data is different. Data has kind of properties of mass and gravity and you know, you have to get it from one place to fit it through pipes and latency as a consideration, things like that. So, how do you do that in the most optimal way and SNAP mirror generally solves many of those problems by being a block level construct that is, you know, the data about it is somewhat opaque. It doesn't really care about the format of it. I can move it easily into an on tap in a virtual machine resident in a cloud elsewhere. I could perhaps stand up a NetApp system in a availability zone or a hyperscale provider, perhaps via something like a Rack Connect and then SNAP mirror into it and then make something like in the case of Rack Spaces, cloud servers, instances available to that data set or perhaps you would look at it in the inverse, the data sets available to the instances. There's, hybrid as in my opinion is still very much an unrealized promise of cloud and part of that's because in an ongoing effort to really standardize what it really means, we think that one of the capabilities we have we've got a lot of opportunity to kind of facilitate the hard work of it which is the movement of the data and accessibility of the data. Cool. So a lot of our customers that are deploying OpenStack are either large service providers or enterprise customers and I guess I'll focus on the networking aspect of the use case, I think for us to kind of solve a problem for them is around how they can get the promise of OpenStack that it already delivers in terms of spinning up new VMs very quickly and connecting them together but have that provisioning work all the way through the end-to-end solution so they can spin up a VM in seconds but their network, their physical network I should say either has to be kind of statically configured or you need to open a work ticket with your networking team so by being able to they're really looking to integrate the provisioning both from the VM side, the virtual switch side and the physical switch side so that when they go and spin up new workloads it really happens end-to-end. I think on the flip side you have the networking teams that are running these cloud deployments that have now lost some visibility into what's going on in their network and they want to be able to correlate the OpenStack the workloads that have been placed around their network especially as you get to overlay technologies where all the traffic between VMs is being tunneled they're kind of losing some visibility into how their network is performing so how can we provide that visibility in an OpenStack environment back to the networking team to give them enhanced debugability and the ability to run their network as they did before in a non-cloud environment in terms of the hybrid question one of the things that certainly we've been working with certain partners on and that we see as a big part of this is encapsulation technologies like VXLAN, at least when we talk about it at the networking layer where how can we take multiple data centers could be on-prem or off in the cloud or just in different physical locations and connect them together and orchestrate things such that you can have VMs that are part of the same virtual network that span multiple physical locations thanks so talking about CMicros use cases and what we're seeing out there so for us a big part of it is how do you manage this converged infrastructure and what we're finding what a lot of our customers want is they want an easy way to deploy OpenStack they want an easy way to manage a pile of servers a pile of storage a pile of networking because really what people want to do is they want to get to value so in a nutshell for us what we find really compelling what our customers find compelling about what we're trying to do with OpenStack is the fact that it makes it easy for them to quickly deploy infrastructures of service and their applications I'll talk about a couple of my favorite use cases that we can talk about publicly one of them is a gaming vendor RedFive Studios they're building a massively multiplayer online game and they're actually using OpenStack to run their infrastructure they're planning to run both in the cloud and they also have a private deployment and in this private deployment they've used their system and one of the specific cases that we help them solve is if any of you go to gaming conferences there's a bus that they're starting to drive to this conference to demonstrate the game and on this bus it opens up and they have a bunch of gaming consoles and they run a bunch of servers so they can run some games but they needed to have the entire system to have the storage, the servers and the networking and we really saw that problem for them because they were able to take one of our systems put it in the bus and deploy the entire OpenStack and their entire software stack onto the solution so it was one box plug it in, one cable coming out and it provided their whole cloud solution for their hosted game moving forward now as we see things evolve how do we get to the hybrid cloud so another use case we have is we just announced this we're doing this together with Rackspace and Rackspace's private cloud edition is for the University of Texas San Antonio and this was for one of their scientific computing groups research groups in the university and they wanted to build a biological research computing stack and one of the issues a lot of customers we've talked to have had with OpenStack is the difficulty of deploying it a lot of people have to invest a lot of their own engineers and a lot of their times try to figure out how to get this stack deployed so it takes a long time to get the value and I think what we're starting to see with some of these supported distributions out there that are supported certified on hardware is now customers instead of having to spend weeks deploying can now try to get that deployed and installed in hours if not a day and what this does is improves their time to value because what you really want to do is you want to get to that utility model and I think that's what us both as vendors as those in the community who are trying to enhance OpenStack make it more enterprise ready and also support it can really bring and help grow this into to the point where it's more adopted by customers and we can really bring some value and accelerate the acceptance of OpenStack Excellent, thanks Thank you, so from a use case perspective I believe that in a nutshell I've got the same read that my colleague here when we engage a prospect or organization on OpenStack what comes across is the application that are really target for OpenStack and their OpenStack management I would say, all right big data type solution but also VDI I also seen some interesting instances where OpenStack wasn't so much looked as a cloud operating system but very much as a way to fully automate the provisioning process so you've got companies out there that are looking at if you will taking OpenStack and sliding under their home application tooling and scripts in such a way that the business intelligence that is included if you will as part of those tools can be leveraged against OpenStack and the value of OpenStack is about automation but it's about also providing this consistent set of API across compute networking and storage that has tremendous value to many organizations out there right purely from on the hybrid discussion I think that there's a an expression, a sentence that is stuck in my head that I heard I think two years ago here at the OpenStack summit and the term has been coined by I don't remember the name but the guy said that you need to own the base and the spike and I felt like it was a very good definition of what a hybrid cloud should be to make sure that it's actually more cost effective to actually run anything that is predictable in terms of traffic into your private cloud and really just offload the spike into public cloud and as a matter of fact when we talk to our customer we actually have that discussion about how to move away from public services and actually bring back their instances to create a private cloud and some of them are definitely considering OpenStack as a way to do this so this industry is basically set on this holy grail approach which is the ability to move VMs between public and private environment and we also have the acquired recently dynamic hops which is basically an automation system that can move application workloads we also believe that workspace is ahead of the curve and that discussion with their private cloud offering and Parquet is definitely part of this offering as well we have had our first round of certification back in January this year and I think that this should come now so we are about at the end I think we've got one question here that will wrap us up really quick we've got about a minute so 10 seconds of peace maybe 10 seconds of peace will be a trick when I look at the members that are represented on the panel there's at least one of you for whom this integration project is kind of a natural part of how you do business for the rest of you probably not so much and so my question is how is being involved with the foundation how is using this as a way to generate value for customers how is it both challenged and how is it changing the way that you look at how it is that you take essentially a hardware or a piece of firm technology value and deliver that out to customers because I think that how some of these how some of the hardware companies overcome that challenge will say a lot about how much value they can provide out into that arena so before we do that you ate up the entire minute we had left with the question so I don't know what the rules are I'm running over I don't know if they're going to hook us like they do with the Oscars so we want to rush through it go so yeah it's certainly a challenging environment for a large hardware company like EMC but we're clearly smart enough to be very excited about the interaction of cloud computing and open source and we've made great headway at EMC to get them focused on this opportunity our goal is to protect the community imagine that's who you're referring to so we've been doing open source development over a long period of time we've derived benefit from it in our own products we've had maintainers of NFS for example and Linux on staff for years is just an outgrowth of it open source is not anathema to us not remotely I would say while we're a hardware company actually a lot of the value that we sell to customers is actually on our software EOS is an open operating system built on top of Linux we push back all of our support and changes to the community and keep up to date with what's going on in Fedora and the Linux kernel this is just like you said another version of this and we're excited about being part of it as we said our focus is really how do we take the standards with NOVA, Swift, Cinder and Quantum and we want to integrate that into our management later to make our hardware easy to use we're actually not trying to change that one area we do want to innovate which is different for the community is bare metal computing and that's what we're working on as my other panelist said we want to protect the community and we want to push that into the community as an open standard so thanks for that question I'm going to be very candid here here at Bokeh we have been a bit challenged in terms of what embracing open source afford I think that this open-site initiative within the company helped a lot for us to realize that there is actually goodness in terms of contributing to open source community and as a matter of fact the vision has completely changed and we recently acquired a company called VRTAT that did a virtual router solution that is basically completely open source based thank you very much thank you panels thank you