The National Aeronautics and Space Administration (NASA) and the National Geospatial-Intelligence Agency (NGA) are moving beyond the earth-bound mission edge and into the stars as they look to launch artificial intelligence (AI) capabilities into space, top agency officials said on Nov. 19 at the Red Hat Government Symposium in Washington.
Federal agencies are looking to send data centers into space and use AI systems in orbit to maximize efficiency, the officials said.
By placing AI systems at the edge in space – such as on satellites, rovers, or lunar systems – more efficient uses of resources like power, bandwidth, and storage can be enabled, said Josiah Johnson, systems architect for the Operations Support Center at NASA.
“Right now, we have to bring things down to earth and then make a decision and then send it back up,” said Johnson. “When you get further and further away, and that latency becomes higher and higher …you can use that valuable bandwidth for other things, like collecting and bringing things back home, the important pieces, instead of just every bit [of] data – that will mean that we can be more efficient.”
Sending data centers into space can also lessen reliance on Earth-based systems, further reducing latency and allowing data to be analyzed and decisions made closer to where data is being collected – such as on satellites or lunar systems – said Chris Johson, the deputy chief technology officer at NGA.
Space-based data centers will allow for real-time data processing and analysis while in orbit – which is different from traditional satellites that collect data passively.
The challenge with increasing capabilities in space, he noted, is securing them and ensuring that they receive as many protections as Earth-based assets.
“Historically, as we have built the satellite infrastructure, it has been mostly protected by kind of being off grid or offline, as it were, from everything else, and there were these degrees of separation,” said Johnson.
“Right as you start talking about putting compute in orbit, that means you’re going to have to be moving code into orbit at a much more frequent rate than you would before … compute in space is going to be iterative and innovative about as rapidly as we’re doing it on the ground,” he continued.
Regular updates and monitoring, along with clear protocols for isolating compromised systems in the event of a cyberattack, helps to build resilience to ensure systems can continue functioning under a threat, Johnson said.
“We have to think about cybersecurity of space assets the exact same way we think about cybersecurity for our ground-based assets,” he said. “Those are all like really, really important aspects of I think this new generation of space-based data centers.”
