NASA Nebula cloud spans nine space centers
Open source floats
NASA's Nebula project – an Amazon-like "infrastructure cloud" for use within the federal government – is now being used across nine NASA centers, including the agency's Jet Propulsion Laboratory and its headquarters in Washington.
According to James Williams – the chief information officer at NASA's Ames Research Center in Silicon Valley, where the Nebula project originated – the service is now providing readily scalable processing power to 300 NASA users spread from one end of the country to another. It's also used by the White House, which provided early support for the project through federal CIO Vivek Kundra.
Nova, the compute engine and fabric controller that underpins Nebula, was open sourced by NASA last year, and it's now part of the OpenStack build-your-own cloud open source project cofounded by NASA and Rackspace. NASA Ames CIO James Williams spoke this morning at the project's regular design summit, held this week in Santa Clara, California.
Officially, Nebula is still in beta. But Williams said that the project will jump through all the federal hoops needed to attain "operational status" within a "couple of months". Williams, who helped drive the creation of the project together with ex-NASA CTO Chris Kemp, says that the service runs on about 2,600 servers and that it has dropped server utilization by about 15 per cent. This can potentially save the company $6m in server costs, he said, even when excluding savings involving power, cooling, networking overhead, and system admin.
The service is used by, among other things, NASA's Servir.net project, which analyzes satellite data and other images to monitor and forecast environmental changes and, eventually, to improve responses to natural disasters. ®
Sponsored: DevOps and continuous delivery