15 Years of OpenStack: How a Mars Satellite Revolutionized the Data Center
In the past 15 years, the cloud computing project has grown steadily. Where does it originate, and is OpenStack still relevant in the age of Kubernetes?
(Image: OpenStack)
On October 21, 2010, the US space agency NASA and the cloud provider Rackspace released the first version of OpenStack under the name “Austin.” The project was originally announced in July 2010 at the Open-Source Convention (OSCON) in Portland and quickly garnered several prominent supporters. However, only two services were on board at the launch of the open-source cloud computing infrastructure: the compute service “Nova” and the object storage service “Swift.” The current OpenStack version “Flamingo” now includes more than 30 different components.
Images from Mars
On August 12, 2005, NASA launched the “Mars-Reconnaissance-Orbiter” with the goal of exploring landing sites for future Mars missions. The spacecraft carried the most advanced telescope camera to date. The camera captured huge amounts of data that NASA's Jet Propulsion Laboratory (JPL) was to make available to the public. According to NASA's CTO at the time, Chris C. Kemp, the size of a file was 600 megabytes.
This presented the space agency with a problem: as part of its research mandate, the data was to be shared with the world as efficiently as possible. However, when collaborating with various organizations, the team at the Ames Research Laboratory encountered an obstacle, as not every cooperation partner could access the required compute infrastructure. Therefore, a solution had to be found to make the data available not just in “Google Earth”.
Ultimately, the team resorted to the Nebula compute infrastructure, which had been developed at NASA for several years. Instead of costly data center rentals, the agency purchased servers and installed them in shipping containers. This expanded use of the Nebula infrastructure necessitated the redevelopment of some components. The space agency decided to release its Python-written cloud computing controller under the Apache 2 license.
From Open Source to OpenStack
The Nebula platform attracted great interest from the American cloud provider Rackspace, not least due to its free software license. The company decided to put the development of its computing controller on hold and instead, together with NASA, to launch an open-source compute infrastructure under the name “OpenStack.” The company contributed its object storage service to the project. The first release, “Austin,” saw the light of day on October 21, 2010.
Even before its release, the project gathered some major supporters who offered their distributions in the following years. Among the best known are HP, IBM, Red Hat, Mirantis, and Canonical. This development led to NASA withdrawing from development in 2012 and instead using the platform as an end-user in the future. In September 2012, the “OpenStack Foundation” was founded was founded to promote the project and the community. In 2020, the foundation announced its rebranding to “OpenInfra Foundation.” At the beginning of 2025, it merged with the Linux Foundation.
Videos by heise
More Alive Than Its Reputation Suggests
According to a survey from 2022, over 300 public cloud data centers use OpenStack. In total, users manage more than 40 million compute cores. The communication service “Line” alone operates an installation with four million cores. The infrastructure platform is also gaining popularity among major European cloud providers such as Deutsche Telekom, StackIT, or OHVCloud. Furthermore, OpenStack plays a crucial role in operating telecommunications networks. For example, AT&T, China Telecom, and Telefonica use the platform to operate and virtualize their networks.
The infrastructure is not only used in various sectors of the private economy but also in many research institutions. For instance, in 2013, the CERN brought its cloud service online. With over three hundred thousand cores, it is one of the largest in Europe. Forschungszentrum JĂĽlich also operates a large installation with the JSC Cloud.
But is the project still relevant in an era of orchestration platforms like Kubernetes? The Open Infrastructure Blueprint from the OpenInfra Foundation provides an answer. In it, the foundation describes the two platforms as complementary technologies. While OpenStack itself provides components for operating Kubernetes clusters, the core services of OpenStack could also be managed on Kubernetes, thus benefiting from advantages such as increased resilience.
(mho)
