Red Hat, Inc. (RHT)
April 15, 2013 12:00 pm ET
Brian Stevens – Chief Technology Officer and Vice President of Engineering
Radhesh Balakrishnan – Head of Virtualization
Hello, everyone. This is Brian Stevens. And I’m here today with Radhesh Balakrishnan, and we’re going to give you an update on Red Hat’s OpenStack strategy, specifically around the vectors of new efforts in the community as well as on the commercial side.
I wanted to start by giving you a little bit of history on Red Hat’s involvement in OpenStack. And it really began back in 2011 in August when we assigned our first developer to start contributing. Although our goals initially were not overly large, they were really how do we take OpenStack, which is really getting a lot of vibrancy and global brand recognition, how do we get that great technology inside of Fedora. And that was our first ambition.
And that followed in August of 2011 with Red Hat redefining our strategy that we were going to fully embrace OpenStack as part of our core cloud capabilities. And at that point in time, as 2011 progressed, we began actively assigning developers because it’s very important that you actually show up to advance the technology and not just distribute it to your customers. And by the Essex release, the first release actually that contain Red Hat code inside of — Red Hat contributed code inside of OpenStack, we have become the #2 developer of OpenStack, something that we remain very proud of.
Later that year in 2011, Rackspace began the discussions around how do they actually take OpenStack and actually federate it and put it into the hands of a governing foundation. And Red Hat became very involved with those discussions with Rackspace and the other industry partners. And then by 2012 or early 2012, the foundation was up and running and Red Hat had joined as a platinum member.
As 2012 progressed, Red Hat became more publicly clear about what our strategy on OpenStack was, in that we were going to create discrete OpenStack distribution and deliver it to customers under subscription. In summer of 2012, we actually made the first preview of that work available for free, and that was based on the technology of Essex.
As that continued and Folsom came out later in 2012, we updated that preview with a Folsom-based release, and then we expanded to create a lighthouse program with up to 25 customers that we had worked shoulder-to-shoulder to help them get OpenStack deployed into their environments. And then that moved us into 2013 with the month of April being the release of Grizzly, the upstream community release, where Red Hat, to our surprise and our great pride, we became the #1 developer of the OpenStack release.
So on the — while it’s really important that the technology advances, it’s key to Red Hat’s strategy that we need to work upstream and that we need to advance the technology. It’s part and parcel with how we think software should be developed in a collaborative model. And so our — the development that we’re doing upstream is not the end, it’s just the means to an end, which is to provide customers with a rich environment around OpenStack. And so the best part about it is as our contributions have increased, the balance between corporate contributions and individuals has also become similarly more balanced. And so you’re no longer seeing one company just solely lead the effort, it’s become a very balanced effort, which I think is great for other companies as well as end users.
And the new announcement today is that when we talk about our life cycle, a vibrant upstream in which we get involved is very important, as well as our subscription-based relationship with our customers. But community distribution plays a key role in that. It’s really a part of our life cycle on how we take upstream technology, get it into the hands of millions, and then subsequently create relationships with tens of thousands of customers under subscription.
So today, we’re announcing RDO. And what RDO is, is an enterprise packaging of OpenStack, upstream OpenStack package for our Fedora and Red Hat Enterprise Linux and other Enterprise Linux rebuilds. And so what it will do is it will track upstream very closely. So as every milestone, every release and ultimately even when key features are released in interim base levels of upstream OpenStack, through RDO we’ll be able to put those into the hands of developers and end users around the world that are using Enterprise Linux. And that capability and that technology is available today at openstack.redhat.com. So that’s the place to go to learn more about RDO and everything else OpenStack at Red Hat.
Early last week, the OpenDaylight consortium was formed, and it’s — and I think it’s the industry’s recognition that the way to achieve mass adoption and interoperability is through the industry aligning around common open source efforts. And so OpenDaylight was the industry’s recognition that the network needs to be federated and that there’s a lot of capability that’s going to be able to be achieved through software-defined networking. So the industry united around Project OpenDaylight that today is being seeded with early open source technologies. And we’re working in a collaborative fashion to determine how to build open SDN-based solutions.
And Red Hat joined that as a platinum sponsor. And the reason we did is because that as important as Linux and KVM have become, as important as OpenStack is becoming, the ability to actually have software-based network control is going to be critically important. And open source is only path to achieve that. So over time, you’ll see, just like we’re announcing today our strategy around OpenStack, over time, we’ll disclose more about how an open SDN fabric fits into our current products.
So now here is Radhesh, and he’s going to bring you through our commercial updates.
Thank you, Brian. Good morning, everybody. Thank you so much for making time to be with us. I’m going to touch upon a couple of interesting announcements that we do want to make today. But before we go there, I do want to touch upon and provide a backdrop of how we think about open source products coming to market.
So if you look at our Red Hat approach, our model, there are 2 parts to it. The first part is engaging deeply and actively in upstream community projects and showing up and making visible contribution thereof. You’ve heard Brian talk about our efforts in OpenStack upstream space and the results to date. And we’re going to keep focusing on that one. The second part of the equation is taking the bits and the innovation that comes out of the community-driven innovation and take it to a point of stability so that enterprises can bet their infrastructure on that. So innate in that is the aspect of working with partners closely to make sure that hardware, software and services capabilities exist so that customers can land a solution and have the peace of mind that it’s not just Red Hat that’s behind the solution but also the ecosystem that’s behind that. So we’re taking the same approach to OpenStack as well.
Our journey when it comes to enterprise products started in August 2012 when we had a Tech Preview of the Essex-based Red Hat OpenStack release announced. We have seen about 4,000 or north of that in terms of downloads of the Tech Preview bits. We’ve also been, like Brian mentioned earlier, engaged deeply with a set of strategic customers as well. There are a couple of things that we have learned from our journey so far. One is that OpenStack is still maturing very rapidly, that it’s not ready for broad-based enterprise adoption yet. We see huge opportunity, but we’ve got to be careful about the use cases and the set of customers that can adopt it today. The second learning is that the most critical dependency for a successful OpenStack implementation is recognition within the IT shop that this is a strategic bet that they’re taking and their willingness to, in fact, invest resources in-house to work with us.
So keeping both of these in mind, today, we are announcing advancement in our life cycle to an early adopter program. So we want to engage north of 100 customers worldwide around Red Hat OpenStack bits based on Folsom release starting today. So these are going to be based on nomination and invitation from our sales force, and we are actually augmenting sales capacity as well as technical capacity around OpenStack in our sales force as we speak.
In addition to making sure that the enterprises have a stable set of bits that they can bet their infrastructure on, it’s also equally important that Red Hat creates an ecosystem around OpenStack to make sure that hardware, software and services partners have a framework to operate with us. So in line with that, today, we are announcing Red Hat OpenStack Cloud Infrastructure Partner Network, targeting hardware, software and services partners. There are going to be 3 elements of the program. Number one is the fact that we provide early access to the bits, as well as have developer support provided to these partners, both hardware and software working with us. The second aspect is actual certification of both ISV solutions, as well as software pieces of integration that we work with our partners. And the third one is around go-to-market activities to make sure that we are driving together with the partner ecosystem adoption of the collective innovation that we’re bringing to market.
So today, if you look at the certification program, what we are designing fundamentally is a test harness, as well as a robust certification program, just like our RHEL certification program, so that partners can engage with us as well as customers can get to a marketplace view of solutions that are today supported as well as will be supported moving forward. So today, we are entering the early adopter stage of this partner program. And I’m happy to say that we have 2 alliance partners, as well as nearly 30 other partners that have signed up to be early adopter partners. And our goal here is to make sure that we get to 10x, 20x over a period of time. So expect to hear more from us on this front.
Talking about alliance partners. The first one that we are excited about is Cisco. As you all are familiar with, Cisco has got a lot of resources to bring to table, and we are excited about this partnership about around OpenStack so that we can provide joint solutions in the market. So expect to hear more on this front as we make this journey.
Intel. It’s needless to say that any hardware innovation needs to be leveraged to its maximum to make sure that we are landing solutions that are enterprise class. So in this context, we are excited about our partnership with Intel, specific to OpenStack. And expect to hear furthermore in the coming months in terms of joint solutions and activities that we’re going to drive in the market.
In addition to having hardware and software partners, it’s also equally important that we have services capability and capacity in the market. So in this context, we are excited about the fact that Mirantis, who are known for OpenStack services offerings today in the market, are partnering with us to make sure that we can scale our engagements broadly in the market. So watch this space closely as we deepen our partnership much more tightly around Red Hat OpenStack and the skill set that already Mirantis has in the form of being able to bring to table today.
And last but not the least, I want to thank the 20-plus partners that have signed up as early adopter partners. And the key point we want to reiterate is that the journey that we’re going to make from now towards the next few months is going to add to the peace-of-mind value proposition that Red Hat always stood behind enterprise products. So a certification program as well as a set of hardware and software partners and services partners is the end design point that we are marching furiously to work.
So just to recap, today, you heard Brian Stevens talk about RDO, which is our community release as well as engagement model, just to further our innovation from an OpenStack perspective. The second announcement is around the early adopter program, where we want to scale the engagement from the 20-odd to 100-plus customers worldwide. And then last but not the least, the cloud infrastructure partner network with 30 partners, including Cisco, Intel and Mirantis already on board, and we are expecting to have even deeper partnership with these partners as well as add further partners in our journey.
So at this point in time, I just want to pause, and then see if you have any questions. And we open the question line now.
So if you have any questions, you can type them into the webcast interface. We have some already. Brian, first for you. Can you summarize the difference between OpenStack.org and RDO, just in summary?
OpenStack.org is really where — it’s really where the upstream of OpenStack is centered. So that’s really where all the knowledge in the community at large for developers comes together. And then RDO, openstack.redhat.com is where we’re at. So we will develop and we will collaborate and submit code at OpenStack.org, as any other developer would. And what’s happening at RDO is we’re basically taking that body of works that’s having upstream and making it so that it’s really easy for anybody that’s running RHEL or other Enterprise Linuxes to actually install it, deploy it, configure it and use it. So it’s really taking that upstream activity and making it instantly consumable to millions of users.
Great. Radhesh, next question is for you. What has the customer response been like to date with the Red Hat OpenStack preview?
So I briefly touched upon this. We’ve had more than 4,000 downloads of Tech Preview. And we’ve also been engaging with “lighthouse” accounts, roughly 20 in count. The primary feedback has been very, very positive. But at the same time, everybody is universally recognizing that from a maturity of the solution, there are still some gaps from an install, configure and management perspective. So we are actively looking at that space to see how we can partner with other players, as well as provide additional value to make it easier for customers to install, configure and manage. Other than that, it’s all about making sure that the core itself is solidifying further and further. And that’s why we’re excited about the Grizzly release. And then our move-forward plan is to get to a life cycle of 3 months from the upstream release being able to have an enterprise product. So the learnings we have from Tech Preview, as well as the further learnings we’re going to get from the early adopter program, are going to inform our strategy.
Great. Next one is also for you, Radhesh. Could you give us a little more color on what types of partners you’re looking for in the partner network?
Yes. So it goes back to the, fundamentally, if you come back from what is going to land in a customer environment, you definitely need storage, network and compute elements to create the fabric. You also need a rich set of ISV applications that the customers can implement and feel safe at night, if you will. So — and most importantly, given the nascency of where OpenStack is, a lot of architecting and services orientation is also in the market. All the demand signals we are seeing indicate that. So that comes down to we need all sorts of partners, i.e. hardware, software, as well as services partners. And we are excited about the level of interactions that we’ve had with the ones that we mentioned, as well as other ones that are underway. So expect to hear more about the traction that we are making in this space.
Great. For Brian, could you explain a little bit about how OpenStack works with other Red Hat products, like OpenShift and Red Hat CloudForms?
Right. Well, you know what, OpenStack is becoming our foundation for building infrastructure clouds. And you can’t do that — while OpenStack provides all the service level and orchestration to build infrastructure cloud, the key ingredient to that remains Red Hat Enterprise Linux with integrated KVM. So that’s where workload runs. So customers are looking for taking their existing applications as well as together with new applications, deploying them on RHEL. The latest data we have is that 72% of all OpenStack implementations are now based on KVM. And so together with the delivery of that capability inside of RHEL, inheriting the ecosystem of RHEL both at the hardware and application level and then bringing OpenStack in together with RHEL to provide — to build an infrastructure cloud is the first step. On top of that comes our Platform-as-a-Service layer, OpenShift. And that sits neatly on top of OpenStack, on top of RHEL. And today, we’re already bringing that to customers in POC environments with OpenStack. And what that means is now for developers that are looking for a more agile development methodology, by deploying OpenShift on top of OpenStack, they get an inherent elasticity for their application workload. And we’ve been doing these proof-of-concepts for some time with customers. And now those are based on OpenStack. And now what CloudForms brings is rich resource management that has to happen across the infrastructure. So today, it’s no longer enough just to build infrastructure and platform clouds. IT needs to be able to have full visibility into all the resources within those clouds, be able to monitor those resources, be able to control who has access to them, be able to deal with things like chargeback. And it has to happen in a very heterogeneous fashion, meaning a consistent view of resources, whether they’re running in the private cloud or whether they’re running in a public cloud. And so together, by ManageIQ now becoming part of our CloudForms brand, we have a rich set of capabilities that sits on top of both our existing virtualization solutions with RHEV but also now on top of OpenStack.
All right. Brian, next question also for you. Could you give a little bit more on what was specifically attractive to Red Hat about the OpenDaylight Project and the goals there?
Yes. What we’ve seen — so it really is really simple. I mean, the networking field is probably the most loaded with acronyms and technology and interoperability issues. But it really comes down to one simple thing is that as you start to build OpenStack clouds, you now — IT now has the ability to control where they put applications and where they put data. And that really matters to them from a service delivery perspective. And they want to move service to where their users of that service are. So now with OpenStack, they have a very agile, fungible infrastructure to be able to move data around and move compute around. The challenge is that the network is inflexible today to be able to support the dynamics they have at the compute and the data tiers. So today, there’s a number of software-defined controllers that are entering this space to try to solve that problem and bring agility to the network layer through software control. And what Daylight serves is, is that the industry recognizing that we need to — the only way to really achieve great interoperability in the network, and it’s critical to not just deliver the function but to be interoperable, is that the industry has to come together and agree on not just the APIs of sort of the old standards’ guards of way of achieving things but actually on the body of code that’s going to form that core controller. And so OpenDaylight to us is something that will be, in the future, a big part of any public, private cloud that we deliver that’s OpenStack-based.
Great. Next question is for you, Radhesh. Could you give any general comments about the timeline for the fully supported version of Red Hat OpenStack?
So the model that we are operating on is we want to get to a rhythm of 3 months from RDO release, Red Hat OpenStack release being available. We have just entered into the early adopter program. The RDO is available today. So first time around, we might be a few weeks here and there. But fundamental thinking is that July timeframe, based on Grizzly bits, we will have Red Hat OpenStack offering, and then from there on get into an N plus 3 months drumbeat.
Okay. And Brian, and building on the response that you just gave around CloudForms, you have a question about how ManageIQ fits into RDO and the OpenStack early adopters programs, if there’s anything further that you could clarify there.
Right. Well, now as of today with CloudForms, with ManageIQ becoming generally available under the CloudForms brand, we definitely have customers and POCs that are happening on the OpenStack side, where ManageIQ, now CloudForms is really bridging — is going that last mile. Because while OpenStack sort of allows them to build these clouds, it still is really difficult at this point in time to be able to monitor and track resources, allocate resources and deal with chargeback. And just about every cloud that we’ve been deploying today on OpenStack requires those features. And there’s a number of key efforts that are happening in the OpenStack community itself, namely some technology around Ceilometer, that’s starting to provide more instrumentation at the OpenStack level. And what that does is that bridges into the CloudForms layer to be able to take advantage of the data — the instrumentation data and present that back holistically. So it’s a key part, it’s an optional part, but it’s a key part of cloud deployment, whether that’s either in the public space or in the private space.
Okay. Great. We don’t have any further questions right now. So if there are any follow-up questions that we could help to handle, if you could email them to firstname.lastname@example.org, we’ll be in touch with the responses. Thank you for joining.