The International Space Station's onboard computer will start chewing on real scientific data, after a year running test suites and benchmarks.

Mitch Wagner, Executive Editor, Light Reading

November 1, 2018

7 Min Read
HPE's 'Spaceborne' Computer Gets Down to Business in Earth Orbit

The Spaceborne Computer on the International Space Station is a little more than a year old, and NASA and Hewlett Packard Enterprise have decided it's time for it to get a job.

The computer was launched to the ISS in August 2017, and its mission until now was to simply test whether it could survive and operate for a full year, the time it can take to travel to Mars. It ran benchmarks and monitored stress levels to test the viability of a commercial off-the-shelf system (COTS) in the harsh conditions of space, including unpredictable radiation and temperatures, power outages, zero gravity, high altitudes and lost connectivity. (See HPE Blasting Spaceborne Supercomputer Into Space.)

Now that the 1 teraFLOP system has passed that test, the Spaceborne Computer will get to work on experimental data for ISS astronauts, Hewlett Packard Enterprise said Thursday.

With limited computing capabilities available in space, many calculations on research projects that start in space are processed on Earth. That works for research on the Moon or in low Earth orbit, between 400 and 1,000 miles above the Earth surface, where communication can be in near real-time with Earth. But when the source of the data is further out -- closer to Mars -- latencies grow up to 20 minutes, requiring local a computer in space, particularly if astronauts are faced with urgent problems.

Figure 1: The International Space Station in 2010 The International Space Station in 2010

Now entering its fifth year, the 2020 Vision Executive Summit is an exclusive meeting of global CSP executives focused on navigating the disruptive forces at work in telecom today. Join us in Lisbon on December 4-6 to meet with fellow experts as we define the future of next-gen communications and how to make it profitable.

The supercomputer project started life in 2014, in meetings between HPE and NASA. The vendor was already supplying NASA with computers to support the ISS, but NASA was looking ahead to a Mars mission, noting that the spaceship would need an onboard supercomputer for that journey.

"We need to know if we are going to be able to take smaller versions of our Earthbound supercomputer to Mars," Mark Fernandez, Americas HPC Technology Officer at HPE and Spaceborne Computer Payload Developer - Software, tells Light Reading.

The Spaceborne Computer is based on the HPE Apollo 40, a purpose-built high performance computing (HPC) platform. HPE picked two-socket Xeon workhorse nodes from the factory floor, and sent them to the ISS with zero hardening. Traditionally, hardware hardening for computer equipment in space is a multi-year, multi-million-dollar process, which results in equipment becoming obsolete by the time it makes the journey, Fernandez says. For example, CPUs on the Hubble Space Telescope+ are nine years old, and the ISS runs 386 processors -- more than 25 years old.

To protect the ISS supercomputer, HPE installed it in a locker in the form of a miniature 19-inch rack. "We do what we call 'hardening with software,'" Fernandez says. HPE uses monitoring sensors on the node with added software to make the device autonomous and self-correcting. The goal is to keep the machine going even in case of failure. Spacebound computers face possible damage from radiation and electrical fluctuations.

"Running as fast as you can is better than slow, running slow is better than idle, and running idle is better than turned off," Fernandez said.

The onboard computer needs to be self-sufficient at least part of the time; the ISS is not in constant contact with Earth. It loses connectivity many times a day, and even when there is a signal, it may be weak or flaky. "This is a very good proxy for going to the Moon, going to Mars, and having a self-sufficient computer that would enable the crews to be self-sufficient explorers," Fernandez said.

Next page: No trains running in space

As a further test of robustness, the computer sat in a SpaceX Dragon spacecraft for a month in the cold environment of space, docked at the ISS waiting to be installed. "We learned that we could package a COTS supercomputer for the shake, rattle and roll and G forces of launch, and for cold storage," Fernandez said.

But the environment in space isn't all harsh. Electricity is free in space, as is cooling -- the machines are water-cooled, and the coolant dissipates for free in space. "It costs zero to operate and zero to cool," Fernandez says.

The HPE computer went up to the ISS in August 2017 and has been operating continuously since activation, except for mandatory power outages, after which it automatically comes back online.

"We've completed our one-year mission and that is all we are contractually obligated to do," Fernandez says. However, the computer isn't returning to Earth right away. "There's not a train coming by every hour to take you home when you're on a NASA space station," he says. The computer is scheduled to come back home in February on a SpaceX spaceship. And until then the computer is going to continue to work.

Figure 2: The Spaceborne Supercomputer has an Earthbound replica. Photo source: HPE. The Spaceborne Supercomputer has an Earthbound replica. Photo source: HPE.

"We're going to open the computer up to space explorers, other users, if they have some onboard processing they would like to do that would give them some advantage," Fernandez says. While most experiments send all their data back to Earth for processing, the HPE computer can process data onboard the ISS and only send back a smaller data payload, making response time faster and precious bandwidth usage more efficient.

The ISS's spacebound computer faces challenges similar to Earthbound IoT applications and edge computing. IoT applications often need to position compute near the IoT devices themselves, to get fast response time and preserve expensive bandwidth. Edge computing devices are often located in austere conditions -- extreme cold and heat, dirt, unreliable power supplies, and vibrations, for example -- that test the robustness of connectivity.

But if the challenges of operating in space can be resolved, space may prove to be a more congenial environment than Earth for data centers. "I jokingly tell people that since electricity is free and cooling is free -- on Earth, those cost more than the purchase of the hardware itself -- as soon as Jeff Bezos can figure out how to get his data center in space, it will be there," Fernandez said.

Spacebound computing, if it proves practical, can have implications for telcos and other communications service providers. Communications satellites today are simply relay switches -- "it takes goes-intas and puts out goes-outtas," Fernandez said. "Suppose we could take the signal in, do some processing, and reduce the output. You've suddenly freed up bandwidth and it costs you nothing." The freed bandwidth "gives you something you can sell," said the HPE man. (See SpaceBelt Secures Cloud With Space Lasers.)

Related posts:

— Mitch Wagner Follow me on Twitter Visit my LinkedIn profile Visit me on Tumblr Follow me on Facebook Executive Editor, Light Reading

About the Author(s)

Mitch Wagner

Executive Editor, Light Reading

San Diego-based Mitch Wagner is many things. As well as being "our guy" on the West Coast (of the US, not Scotland, or anywhere else with indifferent meteorological conditions), he's a husband (to his wife), dissatisfied Democrat, American (so he could be President some day), nonobservant Jew, and science fiction fan. Not necessarily in that order.

He's also one half of a special duo, along with Minnie, who is the co-habitor of the West Coast Bureau and Light Reading's primary chewer of sticks, though she is not the only one on the team who regularly munches on bark.

Wagner, whose previous positions include Editor-in-Chief at Internet Evolution and Executive Editor at InformationWeek, will be responsible for tracking and reporting on developments in Silicon Valley and other US West Coast hotspots of communications technology innovation.

Beats: Software-defined networking (SDN), network functions virtualization (NFV), IP networking, and colored foods (such as 'green rice').

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like