NASA’s JPL deploys private cloud based on Red Hat OpenStack
Relevant links are at the bottom of this post.
The long journey of NASA’s Jet Propulsion Laboratory to the cloud is nearing its end. The Jet Propulsion Laboratory, or JPL, announced this week that it has built a private cloud based on Red Hat’s OpenStack Platform.
JPL’s path toward a unified cloud began in 2008, but it has taken time to develop. That’s partly due to changing cloud technologies and compliance requirements for government cloud acquisitions, but it’s also because of the extreme data demands inherent to JPL.
“It’s very mission oriented. It’s data we’re bringing back via a variety of extra-terrestrial systems. Whether it be from Mars or from satellites monitoring the earth – wherever they are, they are collecting tons and tons of data,” explained Adam Clater, chief cloud architect in Red Hat U.S. Public Sector’s Office of the Chief Technologist, during an interview with FierceGovernmentIT.
With Red Hat’s help, JPL created a private cloud on its own data servers, but because OpenStack ships with Red Hat’s Ceph petabyte-scale storage service, it will also utilize that service for back-end infrastructure, according to Clater.
Initially, JPL went to public cloud providers, but for reasons including costs and control of the data decided “that they wanted to bring that cloud and create a private cloud within the JPL data center to provide the storage and compute capacity to service the needs of the variety of missions that JPL is working on,” Clater explained.
That said, a Red Hat release on the deployment noted that in peak data times, JPL may still resort to public, external cloud resources like Amazon Web Services.
The move comes at a time when the federal government is seeking to consolidate data centers and move government agencies to the cloud. In March, the White House put out its new Data Center Optimization Initiative, which gave new requirements for agencies’ data centers.
But JPL appears to be slightly ahead of the game. JPL and Red Hat worked on implementation of the private cloud for about a year, but have been in conversations about its cloud development for longer, said Clater.
Although the data demands in JPL are high, OpenStack’s open source nature means previous implementations can expand what Red Hat is able to offer, and advise implementations like this one. “We learn a lot from these engagements, but we’re not looking to create ‘snowflakes,'” Clater said.
OpenStack has had deployments in other major research labs in recent years. In 2014, the Broad Institute of Harvard and MIT used OpenStack for a private cloud pilot, and last year Oak Ridge National Lab, a lab within the Energy Department, deployed an OpenStack-based cloud as well.
Obviously, when any organization, particularly a government one, moves its data to the cloud, cybersecurity will be a primary concern. And it’s one that Clater said Red Hat takes to heart.
“As a company, we take security very, very seriously, so we ensure that as our government customers are working with our software, they have a direct line into the businesses so that their security concerns are top of mind within our engineering, product engagement, product marketing, etc.,” Clater explained.
Because OpenStack is run on Red Hat Linux, customers that choose to run OpenStack software inherit the security certifications from its operating system, he added.
– read Red Hat’s press release on the deployment
– read this Information Week post on JPL’s cloud journey
– see this post from the NASA CIO on JPL’s cloud strategy
Halvorsen’s new strategy for data center savings: Cut the people
White House overhauls data center consolidation initiative with focus on optimization
The Case for Partnering