Good day, ladies and gentlemen. Welcome to the Intel Cloud for All conference call. At this time, I would like to turn it over to Mark Miller, who will be moderating today's call. Please go ahead, Mark.
Thank you very much. Good morning. Good afternoon, everyone. Appreciate your time and joining us to have us share our news about our Intel Cloud for All initiative, as well as our collaboration with Rackspace. This afternoon, this morning, we have Diane Bryant and Jason Waxman with us to share some of that information. At the end, we will have brief time for Q&A, so feel free to think of your questions, because at the end you'll be instructed on how to go ahead and ask us our questions. Without further ado, we will hand this on over to Diane.
Okay, thanks, Mark, and thanks to all of you for joining us. We have a couple announcements today that we want to share with you. First is the launch of an initiative to accelerate the move to cloud computing, so accelerating the deployment of new solutions in public clouds, private clouds, and in the hybrid cloud model as well. That's our Cloud for All initiative. The second thing that we want to do this morning is share a joint announcement that we're making with Rackspace, and this is the first announcement within the Cloud for All initiative. We are investing to deploy both the critical developer expertise as well as the lab infrastructure that is needed in order to make our stated objective a reality.
If you think about the environment, the move to cloud computing is happening because there is a very clear and very compelling value proposition. There's value in the cloud to an application developer. There's value to the folks that are responsible for building and operating the data centers, and there's clearly value to the end user that's consuming those services that are being deployed to a cloud environment. The speed to new solutions deployments, the flexibility in addressing customer demand, and that inherent CapEx and OpEx efficiency makes the use of cloud computing technology very compelling. That value proposition that we talk about, it holds equally, it holds across public cloud service providers, for the communications service providers, and for businesses' IT operations. It's a pervasive value proposition. It's why today, about half of all applications are being delivered from a cloud, and that's our estimate.
These are applications that are self-service. They're fully automated. They're living in a multi-tenant and elastic cloud environment. We take that, and we project out that that number is going to grow to up to 85%. 85% of all applications will be deployed to a cloud environment by 2020. As new applications are being developed by businesses, they're going to be developed native to a cloud environment. Despite this compelling value proposition, despite the fact that we've been seeing a 20%-plus growth in the adoption of cloud by large businesses, and despite the fact that most people will say that it's intuitive that cloud is growing like crazy, we conclude that this transition to cloud is actually not happening fast enough. There are real impediments for that. There are impediments for private, public, and hybrid cloud deployments today. I want to talk about them.
The first is the fact that there is a rapid pace of innovation occurring today in the cloud software stack. The result of that pace of innovation is that there are many options for each of the components that are in the stack. From the hypervisors to orchestration components to the developer environments, there are numerous choices and even numerous configuration options within each of those choices. That clearly then leads to complexity in integrating the various software components into a properly functioning, fully functional software cloud stack, a stack that is both reliable and is running at performance. That process takes a high degree of technical expertise today. There's a lot of trial and error that goes into it. A typical deployment takes weeks, at least weeks, if not months, and the solutions are often what we call snowflakes.
They're both unique, and they're a bit fragile. The last obstacle is there are features that are fundamentally missing that an enterprise IT operations organization would simply deem mandatory if they're going to go down the cloud path, such as high availability, such as version control. Management is a big deal. You need to have the management features in the cloud environment that are consistent with the management that exists within enterprise IT today for monitoring, ticketing, authentication systems, all that standard operational infrastructure. With that, our goal is to eliminate these impediments and enable the deployment of tens of thousands of new clouds, both private and public. Success, as we've defined it, is going from an implementation from start to full functional self-service cloud portal in one day. Ultimately, we should be able to do that even within an hour.
To make that happen, to make that a reality, it's going to take addressing the full stack of the software-defined infrastructure. We've got a very simplified diagram here on the slide. There are different layers in the software-defined infrastructure stack. There's the system layer at the bottom, which is all about how you take various compute, network, and storage systems and you create resource pools out of them. There's three boxes at the top that make up the cloud software stack itself. You have the virtualization layer is first, and we use the word virtualization loosely. It could be virtual machines, like running on ESX or KVM, or it could be containers like Docker.
On top of that is orchestration. This is the layer that takes the application and schedules it onto the correct virtualized resources, correct meaning that it's the resource that the application actually matches the demands, whether it's security or performance latency. This layer is also responsible for efficiently managing those resources once deployed in the data center on a real-time basis. This is a highly complex layer in the stack. The top layer is for the application deployment, to allow developers to easily, rapidly deploy their application out into a cloud environment. Given the current state, with the fact that the industry has been making significant investments in this area over the past 5+ years, we believe that what is needed is a very structured approach if we're going to make a difference in the adoption curve of cloud.
What we are announcing then today is a new Intel initiative, that's appropriately named Cloud for All. The objective of the initiative is to accelerate the deployment of highly efficient clouds for broad enterprise market and for the cloud and communication service provider market. A broad, sweeping application. To achieve this objective, we're going to make a range of investments. Those investments you will see can take the form in industry collaborations, financial investments, whether it's M&A or equity investments. You'll see investments in standards development, in market development funds, clearly product launches, and significant contributions directly to that cloud software stack. Specifically, we will invest in the software-defined infrastructure stack to make sure that it is fully functional and easy to deploy.
Second, we will optimize that stack to ensure that the cloud solutions are not just functional, but actually deliver against the expected level of efficiency, the OpEx and CapEx savings that is promised through a move to cloud. Then third, we'll work to align the broad community on standards for ensuring interoperability and portability across the stack. To talk about then one of the very important Cloud for All investments, the one that we're announcing today, I want to hand it over to Jason Waxman, who is the vice president and general manager of our Cloud Platforms Group . Jason, you want to take it?
I will. Thank you very much, Diane. We're excited to be here today to unveil the overall initiative. As Diane pointed out, this is really about moving the industry. Most people look at cloud, they think that the market's on fire, we're in the position to kind of say, "You know what? It's actually not moving fast enough." When we're looking at moving the market, incremental action isn't going to cut it. It's not going to go move the needle. We really need major action. One area in particular that we've seen a high degree of industry interest is in OpenStack. We hear from a number of service providers and also from enterprises that they're interested in OpenStack, they want to deploy OpenStack, again, the challenges that Diane outlined are preventing them from doing so.
In fact, Bloomberg, as an example, is running a ton of services on OpenStack, but they're spending a lot of time adding to the actual source itself to make it enterprise-ready. When we look at that as a major industry need, we thought, "Who should we partner with?" We decided that Rackspace, who is first of all the co-creator of OpenStack, but they're also the company that has the greatest experience of OpenStack operation at scale. We also recognize that between the two companies, between Rackspace and Intel, we held a very common view. That common view is that an open ecosystem will really enable a diverse set of cloud deployments and support industry innovation. That's really the backdrop. With that, we're here to announce today that together we are creating the OpenStack Innovation Center.
Now, the OpenStack Innovation Center is going to have three key components to it. The first is that we want to make the open source around OpenStack enterprise-ready. Together, we're working on a roadmap of features over the next six months to make sure that first we're bringing in things like a better scheduler to make sure that it's going to be more robust, to improve some of the networking capability in OpenStack. To add container services, which has been something that the community's seen in high demand. Just general bug fixes, making sure that the code as a whole is more robust. Now, the impact of this is that today, at best, the most successful OpenStack clouds are seeing scale to maybe a few 100 nodes.
Our goal is to be able to take that and enable enterprise class features at scale of thousands of nodes. Our focus is really on the community and the overall project. One of the things I want to clarify here is that this is not about yet another distribution. We want to make sure that the overall project is healthy and that the people that want to pull from that project have the ability to get the work done that they want to get done. There are a lot of industry partners such as Red Hat, Mirantis, HP, Huawei, and others that are going to benefit from the work that Rackspace and Intel are going to do through the OpenStack Innovation Center. Now, to accomplish this, one of the things we need to do is build the industry's largest OpenStack developer team.
They're going to sit in San Antonio, and part of the reason is that they're going to be right with all the people that are operating that OpenStack cloud at Rackspace at scale. They're going to benefit from that insight. On top of that, Rackspace is going to be upping some of the training programs, not just for the engineers that are in the innovation center, but really to help train the industry more broadly. We're very excited to kind of tap into that expertise. We're going to be adding, as a result, hundreds of engineers to accelerate the overall project. Another companion piece, which we also think is breakthrough, is we've heard the community doesn't have the resources to test at scale. We're prepared, together with Rackspace, to bring online the industry's largest developer cloud for OpenStack development.
This means that anyone who wants to be able to design, build, or operate is going to have a resource that is unprecedented. There will be two 1,000-node clusters that will be brought online within the next six months. We'll actually start to bring some of that capability online even sooner, and we'll continue to make sure we keep the community aware. We really think that this is important to enable people not just to theoretically progress towards their own OpenStack cloud, but to actually build and validate, test, and make sure that the code is robust. Again, we're very excited about this announcement. It's a first bold step forward and a major collaboration for us. With that, Diane, I'd like to turn it back over to you.
Yeah. Thanks, Jason. It is a big announcement that we're very excited about, and the folks on the line may have seen other announcements that we've made or that we participated in over the prior two to three months. All of those announcements have been very well aligned with this overarching, Cloud for All initiative. You've seen announcements that we've made with cloud software leaders like VMware, Microsoft, Red Hat, standards bodies such as Kubernetes Foundation, the Open Container Initiative, OpenStack Foundation. For example, in early May, we announced at CoreOS Fest our collaboration with CoreOS on optimizations for greater efficiency in the Tectonic stack, as well as we announced investments in solutions deployments through our partnership with Redapt. At DockerCon in late June, we announced a collaboration with Docker to improve container security using our Intel Cloud Integrity Technology, and we demonstrated the solution there.
We've taken a steering role in the new Open Container Initiative, which was announced at DockerCon, with the objective of driving consistency in container formats. We joined Google and Docker and others just earlier this week in the announcement of the Cloud Native Computing Foundation. It's targeted at enabling that next generation of application development developed natively for the cloud. With our formalization now today of the Cloud for All initiative, you can all expect that there will be a continued stream of investments that we will be making, and you should expect 15 to 20 major announcements over the coming year. Not surprisingly, we're not the only ones that see the opportunity for market acceleration when it comes to cloud computing.
CoreOS, a developer of open source projects for Linux containers, and Mirantis here, a leader in open source cloud deployments, they clearly see the value in the Cloud for All initiative as a unifying effort for us to drive the market. Between broad industry support, a clear agreement on the problem statement, and the opportunity that follows, we at Intel feel compelled that we need to take this action. We clearly see the business value in accelerating cloud adoption. As you can see here, this is from IDC, and they project that the spending on hardware infrastructure in support of cloud solutions will continue to grow at a 14% CAGR, that's across public and private, out through the foreseeable future.
They state that based on the benefits, the fundamental value proposition of cloud-based computing, the cloud IT infrastructure growth will continue to far outpace the growth of the overall IT infrastructure market well out into the foreseeable future. We believe that we can change the curve through these very targeted investments, targeted clearly at those barriers that we see for adoption today. At the root, cloud computing is another great example of Jevons paradox. It's tried and true. If you make technology more efficient and easier to consume, the rate of adoption will accelerate. I want to thank you all for your time. We certainly look forward to keeping you informed as we continue to invest to accelerate the move to cloud computing, and you'll hear many more announcements in that vein in the future. I'll turn it back over to you now, Mark.
Great. Thank you very much. Thank you, Diane. Thank you, Jason. As we move into the Q&A portion, I just want to do a quick reminder that the press release and a replay of this will be in our newsroom, please feel free to go there for more information. We have links to blogs and Rackspace's news as well. There's a lot more to be had here. Christy, can you help us with the Q&A portion, please?
Yes, certainly. Thank you. Ladies and gentlemen, your question and answer session will now begin. If you do wish to ask a question, please key star, then 1 on your telephone. If you then decide to withdraw your question, simply key star 2. All questions will be answered in the order received, and you will be advised when to ask your question. All of the lines will remain on listen only. Just to remind you, if you do wish to ask a question, please key star, then 1 on your telephone. Thank you.
Thank you. While those questions are queuing up, Jason, Diane, one question that I think has been pretty common is you say tens of thousands of clouds. Can you tell us how you got to that number?
Yeah. I'll take it. It does perhaps seem like a large number, but it's actually not large at all when you think about some simple data points. Today, we estimate that the number of enterprise IT that have actually deployed cloud solutions today is in the low teens, so about 12%-13%. There's 45,000 publicly traded companies today, that tells you there's tens of thousands right there. You can add on to that 100,000 plus managed service providers. You have all the telco communication service providers. Even if you look at the number of VMware sites, they have a half a million customers. There's a half a million folks running data centers that have virtualized their servers. We believe those people clearly see the value proposition of going from a virtualized environment to a cloud-based environment.
You might say that tens of thousands is conservative.
We're bullish. Okay. All right. Are there any questions on the line?
There are, yes. The first question comes from the line of Timothy Prickett Morgan. You're now live in the call. Please go ahead.
Hi, Diane. Hi, Jason. What's the scope of the investment for this? It could be very large. It could be modest. It's hard to gauge. Intel has made-
Yes
pretty big investments. I'm thinking of, like, the investment in Cloudera was quite large, because it's important. What's the scale of money and other resources, because it's not just money?
As you know, we're a pretty big company, and so big for us is big, right? I will say that, first of all, the collaboration with Rackspace, which is the first thing that we're announcing today, is a sizable collaboration. I can say what Diane just highlighted, 15-20 sorts of deals. You're talking about, again, a major commitment from Intel to go invest in accelerating this ecosystem. I'd love to be able to put a number on that for you precisely today. Unfortunately, I-
Me too.
I know. Sorry.
It's bigger than a breadbox.
All right.
Again, Jason, you did explicitly say there'll be hundreds of engineers between ourselves and Rackspace working on the development.
Yeah
thousands of nodes.
You can add up to the thousands of nodes, hundreds of engineers.
That's just one.
training programs and what it's like, and that's one. Exactly.
Got you. Thank you.
Thank you.
Thank you. Your next question comes from the line of Judith Hurwitz. You are now live in the call, Judith. Please go ahead.
Thank you. Got a couple of questions. First of all, the public cloud, is that based on Rackspace? Are you going to offer any of your own cloud services? Are you charging for the use of the developer cloud? The other question I have is what is the scope of the security services and data services within this?
Sure. Obviously, a number of questions. If we leave any of them out, please feel free to follow up. I think the first question was just regarding when we talk about public cloud, we tend to mean the broad industry and the broad market. Obviously, Rackspace does participate in that. The Amazon, Microsoft, Alibaba in China, a whole range of different companies, Salesforce, anyone who provides a service over the Internet, we sort of view into that public cloud bucket. With regard to the developer cloud, our intention is not to charge for the community. This is an investment that the two companies are making to go create a community and then benefit from that growth in the community. We have no intention to go do that.
Security, I think, was the other question.
Sorry.
What data services or security services?
Oh, we as Intel are not planning to be a service provider in that. Again, we benefit as the overall industry grows. We definitely see a lot of collaborators, and we might make additional collaboration announcements with partners in those spaces, but we're not intending to be in them ourselves.
Thank you.
Thank you. Your next question comes from the line of Scott Fulton. You're now live in the call, Scott. Please go ahead.
Yes. Thank you. Hello, Diane, hello, Jason, hello, Timothy, out there. Two questions, if I may. First question having to do-- I don't have to remind Intel about the different classes of processors it produces. The E3 class, which addresses a lot of the needs of the Open Compute Project, the E5 workhorse class. There's the E7 high-performance class, all of which have been said to be innovators in cloud computing space. Then there's the Xeon Phi processors for the super high-performance computing applications. Which of these, or perhaps a mix of them, will we see in the Rackspace Intel community data center?
First of all, I think the straight answer to your question is they will be E5. The other thing which I do want to highlight is when you look across public cloud services, as you pointed out, we have a pretty good handle of where the processors go. Over 95% of the compute and the CPUs that are deployed are the E5 class. One of the things we are seeing is that particularly for high-performance computing in the cloud, we're actually seeing some strong interest and some uptake on Xeon Phi, so expect to see that as being probably one of the fastest growth segments over the next couple of years. Still, by far and away, E5 is the class for cloud.
Very good. Thanks, Jason. The second question, real quick. You mentioned the need to bring better orchestration into OpenStack. Now, OpenStack already has announced Magnum, which is its multi-tenant cloud orchestration, which integrates with Kubernetes. What type of better orchestration can there be for OpenStack than Magnum and Kubernetes?
No, you're absolutely right. First of all, there are Kubernetes, but also Mesos and others. There are some more advanced schedulers that there's some very high interest in potentially integrating together. One of the things that I was really sort of getting on, particularly with this collaboration, is improvements into Nova. It still remains to be the most popular scheduler for OpenStack, and we want to be able to make sure that it can do things like continue to maintain the uptime of services or do things like live migration. These are things that an enterprise is going to require, and we need to enhance that level of scheduler. You're also right on that we're looking at some of those other options in the community.
That also highlights, by the way, the reason that this is so complex, that Diane highlighted, that whole fragmentation and integration of all these different parts together. It's a great question.
Thanks, Jason.
Thank you.
Thank you. You do have one more question. It comes from the line of Brandon Butler. You're now live in the call, Brandon. Please go ahead.
Thank you. You may have touched on this in the previous answer, can you just talk about how Intel products are going to be integrated into the plans for this broad plan that you're outlining here?
Yeah. Absolutely. Obviously, there's a strong motivation on our standpoint is that we want to go see more clouds deployed, we want to see them optimized and taking advantage of all the great features we're trying to expose in the platform. The obvious place is thinking about the CPU portfolio. I highlighted the fact that while E5 continues to be the mainstay of the cloud, we see the emergence of Xeon Phi and other parallel computing. We want to make sure that that software-defined infrastructure is intelligent enough to take advantage of those types of CPUs. It really goes way beyond that for us.
If you look at some of the initiatives that we have around fabrics or some of the initiatives around next-generation memory or solid-state drives, those are all things that a well-optimized cloud can go take advantage of, and we just need to make sure more people have access to that. Even things such as accelerators in an FPGA, an intelligent cloud could really do a better job of taking advantage of that capability. The cloud's got to be optimized and aware of it, otherwise those features fall silent.
Can I ask one follow-up question?
Sure.
Can you just talk about your work that you're doing with major public cloud vendors and how the work that you have been doing with them will support sort of this next stage of work that you're embarking on now?
Sure. We have just a tremendous engagement, and we go very, very deep with all of the large major cloud service providers. A lot of the focus has been around how do we work with them to go optimize infrastructure for their needs. We've literally created a team here at Intel that thinks about every level of what they're trying to do, whether that's the silicon optimization to system design or standards like Open Compute in the industry to rack design, to even facilities design, and then, of course, software optimization along the whole spectrum. We've brought all those capabilities of how to go optimize for the world's largest cloud service providers in the industry.
What we want to be able to continue to do is supporting them the same way that we have by optimizing that technology, but recognizing that there's a whole mass of the next 10,000 clouds that want to be able to do reasonable work to be able to deploy a cloud efficiently and scalably, and we can kind of bring that intelligence to the masses.
Thank you.
Thank you. We do have another question from Judith Hurwitz. You're now live in the call again, Judith. Please go ahead.
Yes. Thank you. You said you're working with all the major public cloud providers, companies like Amazon build their own hardware. Are you including Amazon as a partner in this, or are they going a different direction?
Amazon's a great partner. Again, our goal when we say for all, we really do mean for all. It doesn't mean any menu that everybody wants all the things that we have to offer on the menu, there absolutely is going to be something for everyone, and Amazon is an incredible customer and partner of ours.
Thanks.
Thank you. We have no further questions. I will just do one last reminder. If you did want to ask a question, please key star then one on your telephone. Thank you.
Thank you all again for your time. Thank you, Jason. Thank you, Diane. Just another reminder, there's a lot more information on our newsroom at intel.com, please feel free to go there or reach out to me, Mark Miller, for any additional questions. That should conclude our call today.
Thank you.
Thank you.
Thank you. Ladies and gentlemen, that concludes your conference call for today. You may now disconnect. Thanks for joining, and have a very good day.