OpenStack is going to change everything (in a good way)
As part of their OpenStack RV tour, The nice folks at Rackspace Cloud took a long detour on the way back home and talked to our Cloud Computing Group in Boulder, CO yesterday about OpenStack. In the middle of the talk it finally hit me (quite a bit later than others, it seems) that this is really going to change *everything*. Here's why...
- Rackspace Integration - Sometime this summer, you'll log into your Rackspace Cloud account and see a new (hopefully more intuitive) interface. When that happens, you'll know that the back-end has been officially swapped out to OpenStack, and you'll then have access to cool features like Block Storage (some of which will use SSDs. Hells yea!), Database-as-a-Service (DaaS), Message Queues, and lots of other stuff that AWS has had for years. Rackspace is eating their own dog food and betting their future on OpenStack.
- Maturity - Probably goes without saying, but a major player like Rackspace chucking their proprietary architecture (which was built around CloudKick and SliceHost acquisitions) in favor of OpenStack is a huge signal that the platform is finally ready for prime time. Everything (even RackConnect config) has been re-built around OpenStack and is currently undergoing both internal and external beta test with select customers.
- Interoperability - Can't say it better than the OpenStack Cloud Mission statement: "to produce the ubiquitous Open Source Cloud Computing platform that will meet the needs of public and private clouds regardless of size, by being simple to implement and massively scalable." You'll interface with your private cloud the same way you do your public. Behemoth installs will call the same APIs that your tiny ones do. If you have two hosts who use OpenStack, the only part that will differ is your access creds to each. OpenStack is going to become the no-brainer when a web host or business decides to deploy a Cloud platform. Within a couple months of Rackspace adoption of the platform, there are going to be hundreds of thousands of SysAdmins who know and use the stack, and (assuming the experience is good), they will start deploying it internally, and asking for it in external vendors. If OpenStack gets enough penetration in an organization, closed systems like AWS are going to start getting groans instead of accolades.
- Open Architecture - For the type of system that it is (nascent, rapidly evolving, multi-use, foundational), OpenStack makes sense as nothing else but an Open Source project. While the core team of about 200 developers build the engine & keep the foundation stable & secure, every and anyone else can download the source, hack it up, and/or contribute their piece parts back to the ecosystem. Widely distributed, open source projects focus energy and resources on what's important (what users really want, pushing envelopes, security reviews by the best & brightest) and reduce the chances of that energy being wasted on guesswork and politics. Transparency ensures that security holes, when they appear, are quickly addressed & resolved. My bet is that within a year, OpenStack will offer all of the important features that AWS does -- and after that, it will leapfrog above and beyond, leaving all others in the dust.
- It's Modular & Agnostic - It supports All the major hypervisors, including KVM, Xen, ESXi (with Windows Hyper-V coming soon as well), and the modular, open architecture makes it simple for even hardware vendors like Cisco and F5 to build and provide images and code that allow their hardware to be integrated into/controlled from OpenStack.
- Apache 2 Licensing - All of the above would mean a whole lot less if OpenStack forced the wrong kind of license on it's users. The Apache 2 License it's provided under basically says that anyone can fork the code, build proprietary chunks on top of it, and then sell it back as a commercial product (or keep it open if they like). Cisco and F5 can then charge for those software load balancers, and your company can keep it's proprietary technology proprietary. Brilliant.
As a provider who builds custom Private, Public, and Hybrid Cloud stacks for our Clients, we at quicloud have evaluated (and/or deployed) many of the major engines out there -- vShpere, Eucalyptus, ConVirt, Xen Cloud, and others. Nearly every deployment to date has been an effort in compromise... technologies either didn't work as advertised, or the known limitations, costs, or immaturity kept us from even being able to recommend them to clients.
Finally, I think we've found a clear winner in OpenStack. Am especially stoked to get to play with it as an end-user once Rackspace Cloud launches it this summer. Given all the above, how do you view the future of OpenStack? Is it going to steamroll like we think, or do you feel it's rise will be limited and why? Please drop us your thoughts in the comments below.