While I was putting together my keynote presentation for OpenStack EMEA, I took a trip down memory lane and uncovered documents, emails and slide decks chronicling the earliest iterations of what would later become OpenStack.
It’s been an incredible two and a half years since we launched OpenStack, but let’s take a quick look back and see how we got here.
Most are aware of how Rackspace and NASA joined forces to create an open source cloud operating system. In 2010, the market was looking for alternatives to Amazon’s proprietary public cloud and VMware’s virtualization solutions. Rackspace, as a provider, was looking for scalable technology on which to base our public cloud. At that time, there were very few choices, and the choices that existed weren’t going to meet our needs. There was no choice. All we had were proprietary and/or licensed solutions that would cause customers to be locked into the technology. At Rackspace, our legacy technology was proprietary, too. When we acquired Slicehost to power our first generation of Cloud Servers, we made some code changes, and we had our own APIs and proprietary technology.
We had seen how successful Linux was and we wanted to offer our customers open standards and enable choice. At Rackspace, our intent was that we would differentiate ourselves based on our service and support of those technologies. So we created rCloud – an early iteration of OpenStack, prior to NASA joining. But that original rCloud introductory presentation – the one we shared with folks thinking of joining the initiative – outlined some ambitious goals; some that we’ve already surpassed just two and a half years later.
Looking back now, the summary slide from the rCloud presentation looks like a to-do list: Release our cloud software as open source projects? Check! Run the open source stack in production? Check! Dedicate resources like developers, code and marketing? Check! Check! Check!
And the reasons we did it are still what drive the OpenStack project today. We wanted to accelerate the pace of innovation through collaboration; drive standards through ubiquity to accelerate adoption; eliminate the fear of lock in; open up a closed and fragmented market; and enable public, private and hybrid cloud interoperability with an open stack.
The original note got a great response, but at the same time some folks from NASA released some compute code that also looked really good. We were really excited – here was another company, with a desire to build open source clouds and bring the community to its cause, which seemed to think like we did. It was an organization that had a great brand and major recognition. So, I struck up a conversation with Chris Kemp, then at NASA, now CEO of Nebula.
My email was quick and to the point:
“I run corporate development at Rackspace, and am very interested in talking with your team about Nebula. Confidentially, we are in the process of open sourcing our cloud stack and I am interested in seeing if there might be some synergies/opportunities for the two projects to work together. Would it be possible to setup some time to discuss with your team?”
This is how rCloud evolved into OpenStack – a brief email sent on a Friday morning got the ball rolling. A few days later we got together on the phone, and a few days after that we got together in person. An amazing thing happened during those meetings – as I believe it was during those meetings that the technical meritocracy of our project was born even before it started. Rackspace had Swift code already running in production. The NASA team was building something similar, and rather than argue over who had the better widget, they took our code. And NASA had just started architecting Nova, with only a thousand or so proof of concept lines written. We had started a similar process for compute, but theirs looked good, so we decided to abandon our efforts and use theirs. So OpenStack was born not just with the talking points we all know today, but with actual give and take between the two founders based on technical meritocracy. Amazingly, we agreed on the rest of the principles without debate, and they stuck.
The one thing that didn’t really stick was the name. Deciding on the name was a major debate. I spent more time coming up with names for OpenStack than with the names of my children. rCloud turned into Sangria, CodeRed turned into CloudOS and so on. I think OpenStack had more names than James Bond had love interests. Maybe we should’ve just called it 007. Other ill-fated monikers included Cloudbase, Cloudcontrol, Cloudware, CloudCommander and CloudManager. We even once had an OpenStack logo that had a stack of pancakes.
Luckily we settled on OpenStack, and I think it represents the project very well. I also couldn’t imagine saying that the Rackspace Cloud is powered by CloudCommander. It doesn’t have the same ring to it.
It’s incredible to look back now and see the humble beginnings of a project that turned the IT and cloud worlds on their ears. In just a short time we’ve gone from arguing about names to becoming a full-fledged community run by an independent foundation. The community itself comprises 6,695 people from 87 different countries at time of this blog, and it’s continuing to grow.
Now OpenStack powers not only Rackspace’s public cloud and suite of cloud products, but also our Rackspace Private Cloud Software (Alamo), which delivers an OpenStack-powered private cloud in just minutes. Really, it’s that simple: You can download it for free, enter some IPs and basic credentials, go get a cup of coffee and come back to a fully-configured private cloud.
It’s a step toward true IT-as-a-Service; IT delivered seamlessly that doesn’t require manual updates or upgrades. It just works. You don’t have to worry about the hardware, the host OS, hypervisors or any other component. Alamo is a step in that direction. It’s already been downloaded in roughly 150 countries on all seven continents; by just over a quarter of the Fortune 100 and more than 100 colleges, universities and research centers.
Rackspace Private Cloud Software eliminates the headache of having to worry about dozens of different configurations. We provide an open cloud environment that marries the speed and scalability of our OpenStack-based public cloud with the control and reliability of a private, dedicated cloud environment. This delivers on another one of our promises: the ability to deploy an open cloud anywhere, whether it’s in our data center, your data center or in another data center altogether. It fulfills the vision of having OpenStack everywhere. I like to say “my place, or yours?” It’s in public and private clouds, it’s in your data center, our data centers or in a co-lo facility and you can run like Rackspace and run like an enterprise.
It’s true IT-as-a-Service for customers. Customers deploying OpenStack have the option of having Rackspace support the software and actually run the environment for them, meaning that the same teams that operate our cloud can do the same for yours: monitor it and respond to events; patch it; upgrade it with every new build; plan for capacity additions; deploy best practices; test it and more. Basically, we take care of everything for you. You get an API and an infrastructure that just works. It’s everything you get from a public cloud, but delivered as a private cloud in your data center. You then can truly free resources from working on commodity infrastructure and instead work on the applications that sit on top – the parts that really differentiate your business.
That’s where we are today. It’s incredibly impressive to see how far OpenStack has come since we first started with rCloud, Sangria and a stack of pancakes almost three years ago. I am both proud and humbled to have been involved since the first email was sent, and I’m even more excited about what’s coming in the future.
We bet our company on OpenStack; and I can’t think of a better bet.