One giant leap for edge computing
Humans setting foot on Mars may happen sooner than you think: NASA has its eyes on 2030 for a manned expedition, technology permitting.
Permitting is the key word here. On Mars, computers have to work reliably not only in low gravity, temperatures of minus 81 degrees Fahrenheit, and an atmosphere composed mostly of CO2. They must also be sturdy enough to weather the voyage from and back to Earth, about 35 million miles each way, a round-trip trek that will clock in at roughly two years.
Research suggests that today's supercomputers may be up for the challenge. Case in point: In 2017, HPE partnered with NASA to send an off-the-shelf supercomputer to the International Space Station (ISS), where it withstood zero gravity, radiation, and solar flares to enable an array of astronaut-led research. It operated smoothly for 651 days before returning to Earth.
Building on that success, the HPE Spaceborne Computer-2 will launch to the ISS on Feb. 20., and you can watch a live stream of the launch here at 11 a.m. CT / 5 p.m. GMT. At the heart of the new Spaceborne Computer mission is the data center-level HPE Edgeline Converged Edge system. With twice the computing power of the original Spaceborne Computer, it is the kind of system that could support missions to, and potential colonization of, the moon and Mars. It could also advance research ranging from human biology to climate change, thanks to high-performance edge computing and AI capabilities.
"As we slowly become a multi-planet species, there are new opportunities, and new technologies, every day," says Michael Roberts, interim chief scientist at the International Space Station National Laboratory, who was interviewed for an upcoming episode of HPE's Tech Talk podcast on the Spaceborne Computer-2 mission. "As with any tool, technologies can be used for ill or for good. With Spaceborne-2, we want to embrace the best of these technologies and help them move us forward."
Space: The ultimate edge
On Earth, edge computing is already changing businesses, by collecting, processing, and analyzing data close to where it is generated, rather than sending it to traditional centralized data centers. By 2025, 75 percent of all data created will be processed outside of data centers, according to Gartner. Edge computing not only reduces data processing time and expense but enables technologies such as smart factories and self-driving vehicles by collecting massive amounts of data from IoT sensors and leveraging AI insights in real time.
Please read: Computing on the edge of the final frontier
Space, for now, is the ultimate "edge," providing harsh and remote conditions far from any terrestrial data center while gathering petabytes of data to process and analyze.
"Edge computing is critical. Not only is transmitting 'answers' fundamentally more practical than transmitting raw data, it leverages some of the inherent benefits of the space-based computing environment," says Dylan Taylor, CEO of Voyager Space Holdings. "It is also critical for autonomous functionality in space. In coming years, more on-orbit servicing and space debris cleanup will be required. Identification, rendezvous, and capture will be crucial. This all requires more autonomous intelligence of space-based assets."
Look at the issues for the planned mission to Mars. Data from current probes on the surface of Mars transmit data at a peak of 32,000 bits per second. Sending a mere 250 megabits of data could easily take 20 hours to reach Earth. That's assuming transmission isn't choked by a ferocious dust storm, which can last months at a time on Mars' surface. For comparison's sake, the minimal acceptable download speed for current Internet access to be considered broadband is 25,000,000 bits per second, and users often complain their Internet is slow.
With a facility on Mars being the new edge, local data processing will allow researchers to conduct real-time experiments 35 million miles away (at its closest approach to Earth; at its farthest, it is closer to 250 million miles away) from the nearest data center without waiting days—or months—for algorithmic insights. This isn't just useful; it's essential.
While the first Spaceborne gave astronauts aboard the ISS precious edge computing capabilities, the next Spaceborne Computer will cast a wider net, offering access to research labs, public agencies, private enterprises, and others to conduct space-based research using the cutting-edge system. Research proposals can be submitted online. All proposals will be reviewed by a panel of HPE and ISS National Lab experts.
The most widely beneficial projects will leapfrog to the front of the line, says Mark Fernandez, HPE's principal investigator for Spaceborne Computer-2. "In your submission, you have to state who will benefit from your research," he says. "Is it you, your organization, NASA, space exploration, or humanity? The further out on that list you go, the more likely we're going to welcome you on board."
Medical research beyond gravity's reach
Spaceborne Computer-2 boasts the latest in computing technology, a four-year improvement on its predecessor, offering twice as much compute speed. It's also the first time GPUs have been included in the computing package. The system can ingest and process data from multiple data sources, including rocket sensors, satellites, and high-resolution digital cameras. It can also handle artificial intelligence and machine learning processes along with improved image rendering and other applications that benefit from the inclusion of GPUs.
"We need to send the latest and greatest technology for our exploration," Fernandez says. "If you're about to get on board a rocket ship headed to the moon or Mars, you want to take an iPhone 12, not an iPhone 1."
Beyond analyzing high-res images of, say, the polar ice caps, it will be able to monitor astronauts' physiological conditions for real-time medical diagnoses in space by way of X-rays or ultrasounds. Another potential benefit for astronauts and potential space colonizers is the ability to analyze DNA to monitor the impact of multiyear space travel on a biological and genetic level.
In 2016 and 2017, NASA famously conducted a Twin Study with astronaut Scott Kelly, who spent a year on the ISS while his twin brother, Mark (former astronaut and current U.S. senator), remained earthbound. Findings suggested that a year in space may have accelerated Scott's aging process. Analysis of the twins' genetic data was a slow and cumbersome process, due to latency in data transmission.
Spaceborne Computer-2, however, may unlock faster insights in biomedical discovery in a gravity-free environment, such as its impact on human bone density, immunity, and cognitive function. The ability to parse the data locally and transmit only key findings will allow for a much broader range of studies to be completed.
"Suppose you could monitor the DNA of astronauts weekly or daily, and you could find out exactly when and why things started to change," Fernandez says. "Spaceborne 2 could do that analysis in minutes, rather than the days or weeks it would take to get that data back down to Earth."
Focus on Earth, from a different point of view
While the Spaceborne Computer-2 mission is an opportunity to facilitate state-of-the-art edge computing for researchers with their eyes on the stars, it may prove just as valuable to those studying life closer to home. Hundreds of miles above Earth, in low Earth orbit, the ISS provides a rare window on our planet.
Climate change, in particular, is understandably front of mind for many scientists, as polar ice caps continue to melt at alarming rates, CO2 emissions rise, hurricanes become stronger and more frequent, and 2020's string of devastating wildfires in Australia and California brought entire species to the brink of extinction.
Please read: Using AI to fight wildfires
"From space, we will be able to add samples of data that the [National Oceanic and Atmospheric Administration] uses for weather modeling," Fernandez notes. "That includes visual samples of hurricanes, wildfires, and dust storms that can be rapidly preprocessed, with only the 'good data' sent down to Earth for incorporation into this larger model."
The 4K cameras on the ISS can capture footage of lightning strikes around the clock, but only the 20 seconds before and after the lightning strike may actually be usable data for modeling purposes. With Spaceborne Computer-2, a machine learning algorithm can process gigabytes of ultra-high-def image data in seconds and isolate the useful snippets for transmission to Earth.
It's a turnabout from NASA's typical focus on space exploration, and a welcome one, according to Roberts. "Even though we have a long history of research and technology development that was driven by human exploration of space, this is an opportunity for science that is specifically focused on bringing benefit directly back to Earth," says Roberts of Spaceborne Computer-2.
For Fernandez, the opportunity for Spaceborne Computer-2 to advance humankind's ability to travel farther than it ever has before—and even someday establish permanent human outposts beyond our terrestrial confines—remains a top priority as he considers research projects submitted for the new Spaceborne Computer-2 mission.
"I've got my eyes on the moon and Mars," Fernandez says. "We've got to help humanity make it on those missions, succeed in their goals, and return safely. I don't think we're going to be able to do it without modern supercomputers and GPU-enabled edge systems. Research that directly advances our mission to Mars and beyond, I believe, will bubble up to the top."
As we slowly become a multi-planet species, there are new opportunities, and new technologies, every day.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.