Jo Maitland, senior executive editor, speaks with Pedar Ulander, CMO at Cloud.com, now part of Citrix Systems Inc., to find out what the open source cloud software vendor has planned.
You were VMOps back when virtualization was cool. Now you’re Cloud.com, which is essentially an Infrastructure as a Service (IaaS) platform called stack, that I can use to roll my own cloud internally, or if I'm a service provider I can use it to roll my own service. So when cloud isn't cool anymore, what will you be called?
Pedar Ulander: You know what? I don't think this is one of those things where we're going to see a wave of coolness and then it kind of goes away. Cloud becomes a new operating model. It is a fundamental shift in how people are consuming, delivering and engaging with IT services.
I think while we might see different themes come and go (IaaS, IT as a Service, Platform as a Service, ‘X’ as a Service, etc.), the underlying theme of cloud computing sticks around all the way through. So I think we're OK with the brand. We launched about 14 months ago, and I think having the brand Cloud.com, as well as coming out as a very small company, helped us significantly gain the momentum and traction we have today.
You guys are open source software. Do you have anything to do with the OpenStack open source?
Ulander: We do. OpenStack launched last year, and we were one of the founding members. There were 24 or 25 members ranging from big hardware companies to smaller solution providers like ourselves that got together to really drive open standards and an open community around cloud computing. We've been very engaged in the project since it started.
How are you involved? Say I'm the customer, and I decide I want to buy or use the OpenStack software, and I'm also interested in what you guys are doing. Do [OpenStack and Cloud.com] completely interoperate? If I'm running on OpenStack on Rackspace, could I use a portion [of one or the other]? How does it work?
Ulander: There are a couple ways to look at that. First, OpenStack is a technology community, and it's a place for people to go to download bits and code or engage in a community of users. Most of the time, when you're dealing with a project like this that's very similar to Kernel.org or Eclipse.org, you're going to download bits, and you have the opportunity to build your own OS or you're going to partner with companies who are commercializing that.
For you, we probably wouldn't go to Kernel; we're probably going to go to Red Hat or Novell. The same thing is happening in regard to cloud computing. In many ways, OpenStack is to the Kernel what we are to Red Hat, if you're going to draw a comparison.
There are a few ways we engage with OpenStack on the customer side and on the development side. On the development side, we added Microsoft Hyper-V support for the compute technology. We worked on integrating some of our networking stack into the platform and we're working with them on some API compatibility, so we do have common frameworks for Rackspace, AWS or one of the CloudStack implementations. There's a lot of good developer work happening.
On the customer side, we engage with customers all the time who want pieces of OpenStack in their platforms. We have a couple of customers who are interested in the Swift offering; Swift is the storage service that mimics Amazon S3. We are creating the management framework around that technology so these companies can offer an S3-like service based on OpenStack.
There are service providers who are interested in that?
Ulander: These are service providers on the storage side; I think that's probably where it’s the most relevant. On the nova side, which is the compute piece, we actually work with both enterprises and service providers to really help them on their deployments, whether it's providing support or integration services for their implementation of Nova. Or, you'll see in the future, in a very short term, [that] we'll be able to provide federated management resource pools from our orchestration tool. The user interface will be able to manage a pod that might be based on our CloudStack product and manage a pod that might be based on the OpenStack product. These are things that are coming short term.
And I think a little bit longer down the road, we will ultimately see the two code bases merge. We will be basically taking OpenStack the same way we would take a Kernel and then wrap all our management, optimization and business logic around that.
When you say the code bases will merge, what do I need to take away from that as a developer or IT ops person who's looking at one or either of these stacks? Do I wait for the merge?
Ulander: Most customers don't want to wait, and there's not necessarily a reason to wait. So one of the things we're really focused on in our platform is backward compatibility. As the code matures, it becomes parity with regards to how it operates, how the APIs work and how the calls work. For a developer, it means you're only supporting one platform, in many cases. Or, for an enterprise, they can get started even with our platform and know that their end-goal is to be fully OpenStack. As OpenStack matures, it will become a part of our core distribution.