Nebula, NASA, and OpenStack
This post by Ray O’Brien (Former Nebula Project Manager, NASA Ames Research Center CIO, Acting) was originally posted to nebula.nasa.gov. The Nebula Cloud Computing Platform was the Flagship Initiative for NASA’s first Open Government Plan, and much has been learned, accomplished, and created over the course of this project. Thanks to the team who worked so hard and invested so much in this groundbreaking effort.
When we started Nebula in 2008, it was for a somewhat different purpose than what it eventually became. Under the project name NASA.net, we initially set out to consolidate NASA’s web space onto a unified platform. By offering common web development tools and resources as a centralized service, it was hoped that a convergence effect could be achieved that would result in heightened efficiency and greater visibility into a diverse array of Agency activities.
As with so many great startup stories, in building this system and advocating for its use, it became apparent that flexible infrastructure was a prerequisite to the platform layer. Still with the aim of tackling the web problem, the team took a step back and set out to characterize the state of generic, on-demand, API-driven compute and storage systems. At this point, two interesting things happened. One, we found the state of open systems for enabling cloud-like infrastructure service delivery to be in an early state of development with low overall supported capacity. Second, we noticed considerable opportunity for providing infrastructure as a service in the communities we were courting for the platform layer. With the confluence of these two findings, the team decided to pivot into the infrastructure layer. We eventually came upon the name “Nebula” and set about the construction of an open source compute controller.
Later, on the first open source release of our Nova controller, we found that Rackspace had taken a strikingly similar technical approach to their storage systems and were set to begin the construction of a compute controller just as we were preparing to focus on storage. Given our technical alignment and with the open source release of Rackspace’s Swift storage software, we joined forces to create OpenStack. Our hope was that a community would form around these two pieces of software toward the construction of an open source cloud operating system. To say that our greatest hopes in this regard were met would be an understatement. OpenStack today has the support of hundreds of individuals and organizations around the world, all set on realizing the original vision for the project.
Recently, on May 15, NASA announced a new cloud computing strategy for the Agency at the Uptime Institute’s symposium in Santa Clara, CA. Among its facets is a reduction to our OpenStack development efforts in favor of becoming a “smart consumer” of commercial cloud services. In understanding this shift, it is important to consider that the majority of NASA’s code contributions to OpenStack were during its early stages while its developer and industry communities were still forming. Since those early days, the OpenStack community has grown considerably, and we have borne witness to accelerated development in key areas directly bearing on NASA’s originally identified needs. In fact, the vast majority of code contributions over the past year of intense OpenStack development have come from community members other than NASA.
As the community continues to overtake many of our original internal development objectives, it is appropriate that NASA ramp down its involvement with OpenStack in-line with our announcement at the Uptime Conference. We celebrate this milestone in OpenStack’s development: it has reached a point of self-sustaining growth along a community-driven trajectory such that the project will continue to go forward without our direct involvement. This outcome has always been one of our highest goals for Nebula, and now permits us to transition from the role of developer to that of enthusiastic adopter of a broad range of cloud services, including those based on OpenStack, delivered by a larger number of cloud providers and vendors than existed when the Nebula project was started. NASA has a rich heritage of developing and transferring technology to the private sector for continued commercial development, and OpenStack adds one more stunningly successful entry to that list.
While it is true that our public development activities will no longer be a driving force in OpenStack, should a special requirement be identified by a NASA project wishing to use OpenStack in the future, the possibility of contributing can be explored on a project-by-project basis. That’s the beauty of the community driven open source development model and the key reason Nova, NASA’s contribution to OpenStack, was developed as open source software in the first place.
The announcement should not be interpreted in any way as NASA abandoning or being dissatisfied with OpenStack and its community. Nothing could be further from the truth — NASA is extremely proud of its role in helping to create the OpenStack community. With that success firmly established and many in the industry now likening OpenStack to the “Linux of the cloud,” we are turning our attention to exploring how available cloud solutions of all types can best be used to address a broad set of Agency requirements.