 Ladies and gentlemen, welcome to our panel discussion, the Chief Technology Officers of Cloud. Come on, come on. Yeah, yeah. I didn't know that T stood for technology. I thought it was travel. Good afternoon, everyone. Thank you for being here. We have one more person that's going to join us momentarily here. So my name is Al Sidowski. I'm a Research Director at 451 Research, which is a division of the 451 Group. Some of our sister companies are Uptime Institute and the Yankee Group, and we provide syndicated research and advisory services to vendors and consumers within the digital infrastructure. OpenStack is a coverage area of mine in particular, and we talk to various service providers and software providers. We track the different business models in and around OpenStack, market sizing as well. And we also have an InfoPro service where we talk to enterprises and get their insights, and increasingly they're all talking about OpenStack. So we have a really great panel for you guys today. We have, it's pretty easy. All their titles are CTO. So we have David Lindquist from IBM. Haiying Wang is going to be up here momentarily. John Engates from Rackspace and Lou Tucker. They're all right down the line here. I don't need to press anybody. So we're going to dive in. So one of the first things I just wanted to, you know, let's not do that yet. So I guess guys, the first question we'll ask. Okay, here's Haiying. We've got our fourth panelist. Welcome. So not only is it your job to be technically awesome, but you got to take that technology and, you know, make some money for your company. So I guess the first question for you is, you know, what's your particular, as you go through, we'll just start here with Lou and work on down the line. So what's your company's particular strategy around OpenStack and how are you guys profiting from OpenStack? Sure. Well, I'm a CTO Clack Computing at Cisco Systems. So as you know, we're an infrastructure provider and we provide OpenStack solutions built upon our computing storage and networking as well as services around that, helping our customers build clouds, private, public, large, small. And Johnny Gates from Rackspace and our strategy with OpenStack since the very beginning when we launched it was to have a software platform to power our public cloud aspirations. We have data centers all over the world, including right here in Hong Kong now running OpenStack powering our public cloud. We also have some great solutions for customers around private clouds. We have a number of customers now, you know, large number running private clouds hosted at Rackspace and private clouds in customer premise. And then our real strategy is to put all that together into hybrid cloud solutions that really make the best of cloud for customers. I'm Dave Lindquist, CTO at IBM for our software, our cloud software. Our strategy, and Danny took you through that quite a bit this morning at the keynote, but it's fundamentally is to drive and leverage OpenStack as a ubiquitous de facto infrastructure to service. Earlier this year we announced an open cloud architecture. And at the base of that open cloud architecture is OpenStack. What we're doing is incorporating and using OpenStack across all our cloud deliverables, whether they're the recent deliverables we had this year on smart cloud orchestration, our on-premise software and smart cloud foundations, our integrated systems, our pure systems, as well as our public software. What we're doing is extending the capabilities of OpenStack to address many of the requirements we see coming in from businesses. Certainly the hybrid requirements, as John said, but also the security backup recovery and how to optimize and manage a lot of enterprise workloads on OpenStack. Hi, my name is Hai Yin Wang. I'm from Huawei. So Huawei is a telecom solution provider. It's a leading telecom solution provider. We have been providing IT infrastructure for telco and an emerging market for a while. And then right now we bet on the OpenStack to drive the adoptions in our respect market, specifically telcos and emerging market like China. And OpenStack is a pillar of our solutions that will glue the different pieces we have together as a control panel, tightening together, and a solution. So we really stubble down OpenStack. I believe that's a future to go. Actually, Huawei just became a gold member of the OpenStack Foundation. So welcome. Welcome. Thank you. Great. So 4.5.1 Research did this bottoms-up analysis where we talked to close to 60 vendors that are involved in the OpenStack ecosystem and broke them into six different categories. And there's the PAS players, for example, like AppFog and there's the service providers like RackSpace and IBM. There's DevOps, OpsCode, for example. There's the OpenStack with other clouds. So RightScale, for example. There's the distribution companies who we heard from some today already, Red Hat and SUSE, Canonical. And then there's those IT services, intern key systems solutions, and Morantis would be the example here. And we are estimating 2013, the market to be a little over 600 million, and eclips a billion dollars by 2015. So I guess my question to the panel is like, how do you guys see the business models evolving around OpenStack, moving a little bit from services to products and things like that. So I don't know if somebody wants to take that one first, but how do you see the business models evolving around OpenStack in general? Actually, I think you're underestimating it. You're underestimating the potential here because one of the things we're seeing is there's actually a much broader set of uses for OpenStack as a platform. So we're seeing the cloud service providers naturally, for example, we're working a lot with a lot of the media companies that are using it as the way to run their data center. So I think we're going to start seeing it and people talk about it as an operating system for the data center. It's a new middleware layer, let's say, for these applications, and that to me has even higher potential. One thing I probably should point out, so we didn't include equipment providers, so hardware and software, because it's general purpose, they may be using OpenStack, as the software defined network companies for now, but I appreciate what you're saying. Anybody else care to comment about it? I mean, I guess you're asking how the business model is evolving. I think, obviously, we're covering kind of the obvious things like public cloud, and that's growing rapidly for us. Private clouds hosted at customer premise, and again, hosted at Rackspace, that's maybe not an obvious private cloud kind of deployment model when you think about sort of behind the firewall, that's where most people are thinking about it, but we're also seeing a great number of customers looking at it as a, you know, carved off sort of like a public cloud service. Get it out of my data center, but I still want it to be sort of carved off. I think, you know, embedded in other platforms like a software as a service, we see that, or embedded in a PAS, that's another model for use of OpenStack as an underpinning for those. Every SaaS business and other types, they need infrastructure, and this is just a great way to automate and orchestrate and deploy and manage infrastructure. So it's going to be, like Lou said, it's the operating system of the cloud. It's going to be in a lot of different form factors and locations as we evolve. I think the business models will shift as the value shifts. We're certainly seeing a lot of value currently at the infrastructure layers that are offered by providers or offered through distributions or packaged with solutions, but as the use of cloud shifts from one in which we're getting a lot more efficiency and a lot more dynamics out of composition of resources to one where business units, lines of business can drive new innovative business applications, new business models, new business processes. The business models will shift with that value in supporting these new revenue drivers. Great. So one of the other things we've done is we've talked to a number of enterprises, and we asked them, what are some exciting new technologies and things you're looking at, and when we asked this question in 2012, OpenStack wasn't even mentioned. In 2013, it's now on a list with companies that spend a lot of money on advertising and have products that they're currently in the market, and here's this three-year-old open-source platform that's in the same conversation. So I guess the question, and John, I'll teed up to you as the big service provider right now. So what market segments or use cases are you seeing that are driving some of this adoption? Well, I think, like I maybe spoiled it, but software as a service is one that's just taken off. Every CIO is thinking about getting to the cloud, and I think software as a service plays a key role in getting some of those applications off-premise into the cloud. Lou mentioned media. We see a ton of interest in the media and entertainment space in terms of delivering content, storing content, archiving, all kinds of things in that space. We don't focus on this as a company, but we also have been involved in scientific computing use cases for OpenStack, the guys from CERN are here talking about what they're doing, and their use case, the MIT folks as well. They're on a keynote panel, I believe, tomorrow with Rackspace to talk about how they're using OpenStack in their environment. So scientific and research. What else? I mean, just really the entire spectrum of really internet-facing apps, but now also more of the back-end kind of enterprise apps starting to make their way into the cloud, and as the cloud matures and as the technology set sort of broadens and can take into account some of the needs of those apps, we're seeing a lot more of those moved to... You've probably seen a lot in terms of analytics as well. The whole big data. We're washing data now, and the only way to handle that is on a cloud. Mobile is a huge use case to drive the adoption of cloud computing, and I think all of these sort of go hand-in-hand. It's hard to pull them apart. They're all embedded in one another to some extent. First of all, I think the numbers are showing here. OpenStack is clearly accelerating in the minds of customers, particularly when they think about what's the open platform that they want to base their cloud stacks on. It is accelerating very fast. We recently did a study, had a study of over 800 companies, large and small, across the variety industries, global, with CIOs and CTOs, and the set of the questions were about understanding their adoption of cloud and where they were at, and they quickly broke into three buckets, which was very interesting to look at. Probably well understood buckets, but what was behind it was fascinating. About 50% had embraced cloud and were leveraging cloud within their data centers or public from providers to really fundamentally change the way they offer IT to their organization, so they're getting a lot of efficiency out of the environment. About 30%, unfortunately, I think for that 30%, we're still in a very cautious type of outlook on cloud and we're very conservative in still looking at how they were going to bring cloud computing in. But what was also interesting as the other end of the spectrum, upwards of 20% of the companies were fundamentally using cloud to change their business models. They were using cloud to change the way they interacted with their customers, through mobile and social. They were using cloud to change insights into inventory analytics, into campaigns, into understanding communities, into understanding trending behaviors. They were getting all kinds of decision based insights into how to drive the next wave of business models, business process, business applications. And when you contrasted the financial results of these companies that mapped into those three buckets, you began to see the leaders in industries changing their business models, that 20% with cloud, and then plus the 20% changing the efficiency of their environment, and then basically lagging behind. So it is changing rapidly. We've seen the same thing. It's the business model of our customers that are changing as well. Actually, I think Jeffrey Moore's got a great analogy, which is systems of record versus systems of engagement. And the systems of record with Y2K problem, we probably solved all those. That's done. So your traditional ERP systems and everything else, we don't expect to see a whole lot of growth. Companies are trying to find better ways to interact with their customers, their supply chain, their consumers, and that's driving, I think, a lot of this rethinking about how IT is delivering those capabilities to the business units that are trying to reach out to those very same customers. Yeah, and from what we see in the customer, on the telco and on the emerging market, on the telco, they have a pressure on replacing the current purpose builder infrastructure to support the service that have a social and mobile drive crazy. So we have to move to that direction. And on the emerging market, they have started using ITs, they need to consolidate the infrastructure. So we see the OpenStack have played a unique role here. I think we are all doing great work to support the traditional enterprise IT apps, so the state-of-the-art tightly covered. But in our two sector, we look, they actually don't have that kind of apps. They have different apps. I think OpenStack will be the ideal candidate to lose coupled and scale out. But we're still missing some pieces. So I think that we see the interest almost all the telco providers have a plan to go to the cloud. And OpenStack is really the first choice for them. Great. So some would argue that OpenStack is maybe still a little mature and not enterprise-grade. How would you define enterprise-grade and feel free to challenge the assumption that it's not ready, but what is enterprise-grade and how does OpenStack get to that mass adoption phase based on it being enterprise-grade? So I don't know if you'd like to start. I think enterprise-grade for a lot of years meant it was gold-plated hardware with a badge on it with a support contract behind it and had to be redundant and all of these different characteristics to support that application relied on all that infrastructure to be there guaranteed 100%, right? That's how people thought of it. And I think the definition of enterprise-grade is just dramatically changing because some of the companies that are delivering enterprise-grade software in a SaaS model, they build their software very differently behind the scenes than the enterprise would have imagined or would have thought. We don't necessarily build applications the same way we once did or the definition of what an enterprise expects in a cloud is going to have to change as well. They're going to have to adapt to that and I think the leading edge of enterprises are. I think there's some that still look at it and say it needs to have these characteristics but dramatically, I mean that's dramatically changing and very rapidly. So enterprise-grade I think is almost an old term representing I think the old style of these applications but as I mentioned, as companies that are changing and they're trying to serve, they're not just serving their employees anymore. They're actually trying to reach a much, much larger audience so they're learning from what the web companies have learned a long time ago which is that you need a new architecture and therefore I think, in that realm I think OpenStack is absolutely ready. It's absolutely designed for a scale-out architecture where scaling becomes one of the mission critical things they have to be able to do. They can't afford to miss that spike or that change, the dynamically change flow of interactions that they have with their supply chain. So I think enterprise-grade is actually changing a definition when they start to look outside of just serving their employees with the traditional apps. Haiying, what about from Rawways? I agree with what the guy says. I think the apps would decide what is enterprise-grade. So the early terms are ERP, Oracle-based and the traditional infrastructure or even VMware infrastructure is better because they're more tight and statorful and when you have a look at the new apps, the new SaaS model, they're more scale-out, web apps and OpenStack will perfectly fit so that would be their enterprise-grade. So I think it depends on who you're asking what app you're specifically talking about. Some of the apps are the enterprise-grade and we just need them all. I think that may touch the point that we need the past to be fast and develop more apps for enterprise because the current app is still an older generation. Same workload that will work well for the old infrastructure but it may not be good for new infrastructure. Dave? Well, I think all of us are using OpenStack. I think much of the audience, the companies they are using OpenStack, I do think it's ready. It has been ready. I think the uptake that we're seeing since Grizzly has been fantastic. I think we have, as a community, work to do. We've done a lot of work in areas of security. We've done a lot of work in areas of testing, getting the quality, the availability. I think we do have to focus more on interoperability, interop, particularly as we drive this where everybody wants to take it into a much more of a hybrid environment. But to me, it's clearly ready. I think there's maybe some areas where we could do more. We could add more in terms of some of the security features that some enterprises are looking for, granular permissions about who can do what in the system, how the system trusts other elements of the system. I think that's an aspect of it. It's not so much the enterprise grade. It's enterprise features that they have come to expect that they want to have similar functionality in OpenStack. It doesn't mean that OpenStack can't be used in an enterprise. It just may be used more broadly in an enterprise if we give them a few more capabilities and features that orient it towards some of the historic and legacy workload type context. I actually don't think that the customers we're talking to are not looking to just have applications. It's not that that's happening. They're seeing a whole new class of applications they're now struggling with how do they bring them out and the cloud is the best way to do it. We even see in the telco space we're we're function virtualization. I want load balancing as a service. I want firewall as a service. What they really are talking about is they need those things to dynamically scale. That market is growing so fast in terms of demand that your traditional approach just isn't going to work anymore. So they need to have these new ways of thinking designing applications and that's really what OpenStack is. If anything I think that OpenStack is moving into a lot of these different areas at the same time and as a community we're trying to make sure that we balance the innovation the addition of new services and everything else with a need to continually go back and making it better for operators making it easier to manage and secure. Those two themes are what you're I think you're going to see the community really focusing on. Right. Now compute storage have evolved a lot faster than networking and understatement I guess. Now as far as software defined networking and network function virtualization there's a lot of discussion happening there. Lou I'll let you tee this one up first. We can all comment on it as well but these two things seem to be happening at the same time. I mean Martin Casado had really nice talk about how virtualization now is being applied into network and I really agree that we're seeing that this whole trend around SDN and network virtualization is happening and cloud computing is happening and what we've seen is almost every provider of new SDN controllers and everything else is using OpenStack as a vehicle to have that expressed in their data center. These are happening as sort of two layers in the stack and both are evolving a co-evolving actually right now at the same pace and that is a marriage made in heaven because now the individual applications or the tenant whatever gets to take it gets to be able to have these virtualized data centers that they're deploying their applications on as we move forward into pass and other kinds of things and have the infrastructure now respond accordingly whereas if we had simply a fixed infrastructure it would be very hard to do that. I think Lou is exactly right. What we're seeing in the software defined networking is really pushing the envelope on compute networking as well as various aspects of storage. It's really bringing that layer to an abstraction of a software defined environment where you can very rapidly compose the infrastructure into patterns to meet the needs of a specific workload whether it's scale availability zoning etc. And what's driving that is from a business perspective how quickly can I put new applications or processes into market. How fast can I take what's the platform layer and compose services so that I can rapidly develop my application, get it in field and see if I'm achieving that business outcome. That's dependent on the infrastructure layer becoming a very dynamic fabric through this software defined environment. In OpenStack, the core projects in OpenStack are really the fundamental projects that the industry is looking at to make that software defined whether you call it software defined environment or infrastructure but that software defined layer. But it's the businesses that are driving down because that's how they're driving revenue and their businesses is really to really extend their reach out to their customers through mobile to create very compelling interactions with social to get business insights through analytics and big data and what's happening with the communities what's happening with individuals, what new opportunities exist so that time to market that value is what's driving those compositional needs and to me that's what the core was happening with software defined network, storage and compute. I think the beautiful thing about OpenStack is that there is really no better place right now than OpenStack in this community to do these kind of network transitions or these innovations that we have to do. I mean this is a place where you can have sort of a many to many relationship amongst the network providers, the network vendors, the users, the people that are building large major public clouds and also private clouds. I mean I don't think we could have gone as fast as we've gone in software defined networking if it was just a purely proprietary one to one kind of relationship vendor to customer. I think it really is now set up nicely for this to go faster and faster because there's going to be a pull from guys like Rackspace there's going to be a tremendous pull to make sure these networks can scale they have the features that the customers are looking for and like you said they can be expressed within OpenStack where the customer can actually compose the network however they want to so it has to be flexible, it has to be scalable, it has to be reliable, all those things are going to be just tremendous pressures and pulls on those network. You know as a fundamental principle that we in the community developed around OpenStack it is a set of services and that's what makes it up that are basically loosely coupled but we specifically designed so the specialists in that area can really go and I wish that one service as fast we were part of the very early discussions of what was then called Quantum which was trying to pull the networking out of NOVA because we felt that trying to do networking was just like all mixed up inside of compute and then that meant we could never take advantage of the innovation that's happening in software to find networking and so we really pulled that out and David you mentioned about we were trying most particularly to pay attention to the abstractions if we could get the abstractions right that a developer, an application developer found useful then we have the freedom to explore all different varieties of software to find networking and different things that can be done in the infrastructure and to provide this capability so it allows innovation both sort of above and below the line in OpenStack I think that OpenStack really bring a very good platform to drive network virtualization and storage so server virtualizing did a good job and then we moved to these two sectors and storage fundamentally is too expensive and so people are doing those virtualization and network is just too complicated but I think that the real drive would be the use case applications and OpenStack given is pervasive everywhere, lots of people trying to use different ways and innovation really happen in this place so Niseria is a good example, they use OpenStack as a platform reach out and they're doing a good job I would expect more in our sectors in the telcos we see the workload especially on network environment when you move to OpenStack the big challenge is how you can guarantee that workload that network connectivity they want they're more connective centric than the computer centric so that will bring lots of push on there and I think we expect that in the next few years we see lots of innovation in this space so when we talk to enterprises, when we talk to vendors one of the consistent responses as to why OpenStack is around the community and the vibrant community and one of the things that people question is so when you talk about Amazon Web Services or VMware or some other companies they have a defined strategy the company is all on board in the case of OpenStack there's a board there's a community everybody doesn't have the same interests and it seems like everybody's you know in fact I think we have the same interest what's interesting maybe this period of history or whatever that we see there's a lot of change going on and that we all have customers who are trying to adapt to that change having multiple interests versus having that single is that a help or a hindrance going forward as more and more companies start to profit from the platform. Right now I mean particularly sitting on the foundation board it's been very rewarding because we actually are all we have a common goal and that is to make OpenStack succeed and to make OpenStack grow and we are all dedicated to that goal I think I mean in this last release there are over 900 individual contributors to Havana we've got over 12,000 people in the community today that's a lot of talent and generally the talent is coming from some of the best and brightest from each of those companies so it's sort of like a very elite and in those who are OpenStack developers here I applaud you because I think they really are focusing on and by having these separate projects you're able to get in and you're able to contribute in an area that you really know a lot about so right now maybe it's just a honeymoon period we're in right now but I think that there's every reason to believe that this is going to just continue to grow as we're working together and then in the market we'll fight it out over you know our customers and everything else. I think the other aspect of OpenStack that makes this not as much of a problem is that it's truly a meritocracy in terms of just contributing code. Who's pulling? Who's actively making this thing go forward and I think that's what a community does they get behind an idea they rally behind an effort that they see a lot of promise in and they all pull in the same direction and I don't think that if you perceive debate from the outside looking in I think that may be amongst maybe people like us but down in the developer trenches where the real work gets done I don't think that happens as much it's like who is contributing code who is the smart guy in the room coming up with the best ideas and driving this thing forward. Well code talks Lou mentioned it earlier to me part of the growth the explosion that's occurred in the community obviously the open ecosystem that John just mentioned and the way that's run but because of the way OpenStack was architected from the beginning as a very clean separation of the core domains with well defined plug in extension points and that allowed a community to have a common goal of driving OpenStack to a point of a de facto infrastructure as a service and platform the all of us partner on solutions and compete aggressively in areas and collaborate in areas the contention if you want to call it contention sometimes in a competitive spirit of innovation about doing each other is healthy that drives us to drive things that you mentioned on software defined networking it drives the new levels that are occurring in storage not just in file block object but in tiering and management it drives and compute to not think about just the virtualization but the abstraction and the how do I deploy a complex pattern using a network so to me the competitive energy in the community at large is driving the innovation around these core projects around these well defined points of extensibility which keep a common goal how do we drive OpenStack and then how do we start bringing innovation into that that's very healthy and I think that as long as we continue the community and the common goal and don't fragment which I see that it's consistent coming together it's going to be very strong and healthy I mean imagine if a company, I mean our PTL's are elected by their peers review is reviews are required and before code is committed in Tubstream all of our guys are upstairs on the second floor this week on the design summit for the next release make debating how much is too much what do we really need to get done during this next period of time imagine if you had a company that was organized that way I think that you probably would see the speed and innovation happening much much faster so in fact I think that meritocracy is what is really attribute to the reason why that we're seeing this kind of explosion and I want to give you a chance to get in here as the newest old member of the foundation we're excited to be going I think what we like and we see at this stage at least it's a developer driven and this is probably the most modern software project ever and there's lots of technology going on you track the best people coming to work there and it's basically it's not a solution it's a platform so there's not much business interest that you can directly use that so that's good and we're also in the early stage of cloud development we're clear to target Amazon and VMware so it's a long way to go before you're fighting each other but at the same time if we can keep this the pattern and we can I hope the board can continue to do that we focus on how to get this technology going we have a clear mission we want to be a pervasive open source platform if we're just doing that and I think it will be in the right track and just like in Linux we don't have a problem we just go everywhere everybody making money by doing service and anything else you first make them a best platform first before anything else can happen great well gentlemen thank you I think we're at a time at this point but I hopefully we'll stay in touch and be back here next year talking about the more successes and exceeding the billion dollar revenue targets that are underestimated so I appreciate everybody's in the audience's time thank you