OpenStack stands out among the panoply of cloud options today: The software platform for running IT infrastructure is open source, so developers from around the globe add to and improve the still-young technology. So do huge companies with lots of research-and-development power behind them -- Red Hat, Cisco, IBM, Oracle, Mirantis. The list goes on.
Other OpenStack benefits: The platform is shared, so there are standards for how it's delivered -- a rarity in the largely unregulated cloud realm. And organizations that use OpenStack -- often to build their own private clouds -- don't have to worry about being locked into using one vendor. They can mix and match.
"So you would have greater vendor choice if you decided to switch products," said Lauren Nelson, an analyst at Forrester Research.
Nelson, who helps organizations make decisions on private cloud computing, recently spoke to SearchCIO about what CIOs need to keep in mind when assessing OpenStack benefits. Cost, for example, isn't one of them.
"If it's listed as open source, it's assumed that it's less expensive." It's not, she said. Vendor distributions, or distros, of OpenStack can get pricy -- and while there's a free version, the costs of hiring and keeping skilled staff to maintain the platform outstrips what many organizations can afford.
Cost is just one question CIOs will need to consider during their OpenStack evaluations, Nelson said. Another is more basic: What do they want to get out of private cloud? OpenStack may be the way to go for infrastructure -- but the private cloud market is changing, Nelson said, with development capabilities gaining more attention.
So CIOs might want to look at a platform-as-a-service offering like Cloud Foundry or OpenShift as well and build that on top of OpenStack. And they'll need to think about containers, which organizations are using today in greater numbers to standardize and speed up software development.
Here are excerpts from SearchCIO's conversation with Nelson.
There's a lot of talk about promised OpenStack benefits. What are they?
Lauren Nelson: People talk about the community and how essentially you get free innovative power from all these other companies -- similar roles, similar processes and projects. Typically you have to pay pretty expensive membership fees in order to take part in a community like that. With open source, that's the exception.
The other thing is it's shared, so it's become a standard. There are similarities between this platform and almost any of the other solutions out there. So you would have greater vendor choice if you decided to switch products. You'd also have consistency between others that might have been built on top of OpenStack.
Cost is a little bit of a false one. It's actually quite expensive. When most enterprises think about open source, they think, "OK, some kid in their basement developed that" -- and it's not the case with OpenStack. If you look at the long list of vendors that have fully paid professionals -- all they do is write code for OpenStack -- you start to look at how you're really getting a corporate-funded project across multiple organizations that are building for consistency.
But most organizations are not consuming the pure OpenStack offering -- they're consuming a distro of it. So they're buying a packaged solution like Red Hat cloud infrastructure. And they're looking at Mirantis or looking at IBM. There are a number of solutions out there that end up escalating costs quite significantly.
What questions should CIOs interested in OpenStack ask themselves?
Nelson: The first question they should ask is, 'What's our approach? How are we going to go toward OpenStack?' Because OpenStack differs greatly depending on how you consume it. Are they going to invest in their current staff and try and go with the pure code -- and keep costs at a minimum for the actual software costs itself and invest in training their personnel? Or invest in managed services support from something like Rackspace that will have remotely managed private cloud services on your own premises?
Those two approaches are very different ways of spending money and very different investments. Or they can take a very packaged, very expensive software solution -- and it's in the multimillion-dollar range -- and try to build a private cloud out of it. And that's just the software alone -- that's not even the infrastructure that you need to support this.
So what's the role of private cloud in your organization? Are you investing in it as, 'This is what our entire virtualized environment is going to live in -- and private cloud is our on-premises environment?' Or, 'Is private cloud going to be used like we would use public cloud -- for provisioning net-new workloads or building applications that would be the future of our organization?'
It's less of a 'Do you do OpenStack?' or 'Do you not do OpenStack?' It's, 'Let's define what our private cloud strategy is': What's the cost that you want to spend? What's the purpose? Is it small in scope? Is it very large in scope? And then what's the role of private cloud versus public? And what's your faith in current skill sets in terms of developers and operators of your organization?
And once you come up with the answers?
Nelson: If you're taking a traditional private cloud approach, OpenStack is one of your only options today. There's a costly version and a less cost-intensive version. But private cloud is a lot more than that -- it's about developer audience and scope. It's about talented developers. Containers typically is the approach they take instead of OpenStack. Sometimes they work those together. Same thing with Cloud Foundry: If developers are not your strength but something you need in order to consistently develop code faster -- more of a factory-type approach to development -- Cloud Foundry might be the approach, and OpenStack might be one of the commodity solutions that sit beneath that world.
And [OpenStack has] become the de facto for a lot of service providers and a lot of telcos, a lot of high tech, a lot of academia. If someone is taking a very generic approach to private cloud, [OpenStack] tends to be one of the major solutions that they look at. If they really are trying to look at it from a developer and innovative standpoint and be at the front end of the market, it's no longer about OpenStack -- that might be an underlying one, but it really is containers and it really is Cloud Foundry or an OpenShift that takes a mix of those two approaches.
How can CIOs choose from among the different OpenStack vendors?
Nelson: Sometimes it's what existing partnerships [the vendor has]. A lot of times cost is a huge influencer. [Organizations] want to know what customers are already using you. What size environments are you used to supporting? The user interfaces -- so comparing the user interfaces between the different vendors, capabilities that are built in or incorporated with. One of the big qualifiers is what do you do for storage and networking? Those are two of the bigger pain points in OpenStack, and so which of those distros might align to the storage solution that you're looking for?
Containers give new lifecycle management to OpenStack
OpenStack, VMware duke it out in the private cloud ring
What's in an OpenStack release name?