OpenStack Aims to Please the 'Open' Cloud Crowd
Jim Curry, general manager of Rackspace Cloud Builders, is having a good time giving the people what they want.
Virtualization Review Editor in Chief Bruce Hoard recently interviewed Jim Curry, general manager of Rackspace Cloud Builders. They discussed topics that are on the minds of current and potential cloud users.
VR: What's the current state of the OpenStack project?
Curry: I guess if I were to try to sum it up in one word, it's fast-paced. This project now is more than a year old. During that year, we merged two projects that Rackspace and NASA had been doing separately into a single project. We've been through three releases, and now we're into sort of a pre-production release. I expect that the theme is going to switch very rapidly in the next six months to user stories -- people talking about deploying the project and what works and what doesn't, and giving input to it. We're very excited about that.
VR: What's the most common question you get from CIOs who are interested in OpenStack clouds, and how do you answer it?
Curry: The most common questions we get into are: Where does the OpenStack fit in the realm of cloud? How do we define exactly what it does? What does it potentially replace or what does it potentially add to their existing framework of technologies?
"Cloud Builders' goal, long-term, is to help people support and maintain their OpenStack clouds. Ideally that's something that we're going to do through partners as well as directly with customers."
Jim Curry, General Manager, Rackspace Cloud Builders
VR: What's your reaction when you see a survey wherein a high percentage of respondents say they're either in the experimentation stage or holding off on cloud
projects altogether because the technology is still too immature?
Curry: I don't think it's an unfair comment. I think the cloud is a relatively young technology. I think you have to parse it into two comments. Public cloud is something, I think, that's relatively young.
There are only a couple of providers -- Rackspace and Amazon in particular -- who have achieved any sort of scale in cloud. I think CIOs are trying to understand what it means to consume resources outside of their enterprise, and all the different control issues, compliance issues and resource-management issues that go along with that.
VR: What is Cloud Builders?
Curry: Cloud Builders' goal, long-term, is to help people support and maintain their OpenStack clouds. Ideally that's something that we're going to do through partners as well as directly with customers. It's very similar to the existing Rackspace business and hosting. Customers stand up environments with this, whether they be based on the LAMP stack or Windows environments, and that contains a number of different technologies that we simply make work for the customers and assure it stays up and running. When it comes to something like OpenStack, we're going to do the exact same thing. We're going to help customers get the technology deployed, and then help support them. And then with partners, we're going to actively work with them to make sure that as they roll out technologies, they get the support they need to get deployed into the enterprise.
VR: How would you describe the response you're getting from organizations that are interested in Storage as a Service powered by OpenStack object storage?
Curry: One of the nice things about our project is when we launched it, our object-storage solution -- whose project name is Swift -- was already very mature. It had been in production at Rackspace for about 18 months at that point. It certainly continues to evolve as part of OpenStack, but it was very production-ready at the time, and we'd seen very rapid adoption. NASA took it very quickly and started utilizing it, and we've seen a number of different companies and distributors take the technology and deploy it.
VR: How closely does Cloud Builders work with customers when you build a cloud for them? Meaning, do they learn enough to build their next one without you?
Curry: Yes, and in fact, to a certain extent, we try to make that happen. We work off of published reference architectures, which we're going to make available to the public so folks will see exactly what we're deploying. Part of our effort toward getting deployments done is to actually train the customers on them. So we train the systems administrators on how to deploy and run the cloud. We train them on how to interact in the open source project. We get them up to speed on the technology, and then we stand behind them and provide support.
VR: How do you determine how much to charge customers, and how can they figure out if they're moving toward a reduced TCO or a short-term ROI?
Curry: We're in the early stages of defining our pricing model. We're very focused on making this a solution that's cost-effective versus other alternatives for providers. We're working with what I would call -- I hesitate to call them beta customers, because we're working on a very production-scale offer -- but we're refining what our long-term pricing model could be. I can tell you that the early feedback we're getting from customers indicates that we have a very compelling pricing offer.
VR: Can you give me some examples of your support?
Curry: The best way to sum it up is, when something goes wrong, we're there to help. We're not about trying to define whether it's our problem, whether it's a hardware manufacturer's problem or whether it's another party's problem. Ultimately, if someone's application goes down because it's resting on top of an OpenStack infrastructure, that's something that they want resolved, and we want to be a partner in resolving that. We're partners with customers. They typically know the folks they're working with by name, whether it be the deployment engineers or their account managers after the fact.
One thing to clarify: We're not launching a full-managed-support offer. We don't offer support yet for the guest OS and above. We only offer support from, basically, cement up to guest OS; but longer-term, we certainly are looking for ways we can provide a full-managed-support experience for customers.
VR: Is there one misperception about the cloud that you'd like to dispel?
Curry: I think the security discussion is very common with private cloud, with public cloud and general reliability. So I'm going to take your question and break it in two. I do think that if you work with folks that understand cloud architectures, you can make a very secure offering in the cloud, whether that be in your own datacenters or using public clouds -- or doing both. I think in regard to availability, there's been a lot of hype about public cloud failures and what those look like. I think if you were to look at the world and look at all the different IT environments and their failure rates, it's safe to say that failure rates for privately managed IT environments running on dedicated gears are definitely higher than what you'd typically see in the cloud, and obviously we've had a lot of experience with managing customer environments over the last 10 years.
VR: What's your final comment on security?
Curry: I do think the cloud experiences failures, but it's built to be resilient. I think that once you design an appropriate architecture for the cloud, you can have an architecture that's much more resilient than you'd find in traditional architectures. That's one of the things we're focused on with Cloud Builders: Making sure we can not only help customers build secure clouds, but secure clouds that have a higher availability and reliability than they'd typically see in their traditional architectures.
Bruce Hoard is the new editor of Virtualization Review. Prior to taking this post, he was founding editor of Network World and spent 20 years as a freelance writer and editor in the IT industry.