SpaceX will deliver the HPE supercomputer to the ISS. How will this speed up the mission to Mars?

    Many of you watched the launch of the SpaceX CRS-12 rocket of the company Ilona Mask, which sent the Dragon spaceship to the ISS yesterday. We watched it with special trepidation, because on board the Supercomputer Spaceborne, created by HPE.

    This supercomputer is part of a joint NASA-HPE experiment that will include the first ever test of a commercially available high-performance system in space. The purpose of the system is to function without interruption in difficult conditions during the year, that is, a little longer than the flight to Mars takes. Under the cut, more details about what we launched into space.

    What is unique about Spaceborne?

    To create a high-performance system capable of long periods of continuous operation, it is necessary to increase the resilience of technologies to external factors. The Spaceborne computer, created as part of the current mission, is based on HPE Apollo 40 class systems with a high-speed switching network, and runs on Linux. Although the hardware components of the system were not modified, the company designed a unique water-cooled container, and also developed special system software taking into account the limitations of the external environment and increased requirements for the reliability of calculations.

    According to standard practice, for NASA to approve equipment for use in space, it is made physically more stable. There are enough threats in space: radiation, solar flares, subatomic particles, micrometeorites, instability of power supply and cooling. Physical strengthening takes time and money, and also adds weight. By the time of launch, computers have time to become obsolete for several generations, moreover, they are sent to space after several years of use.

    Instead, HPE engineers took the lead in securing with system software that will manage real-time debugging of computer systems, as well as mitigate the effects of errors caused by environmental conditions. Even without further strengthening, the system successfully passed 146 certifications and safety tests so that NASA allows it to be used in space.

    “Cosmonauts will not be system administrators,” says Dr. Eng Lim Guo, HPE SGI Technical Director and one of the experiment leaders. Spaceborne Computer is housed in a container that will provide it with almost autonomous functioning. It is secured with standard NASA mounts, connected using standard Ethernet cables and 110-volt power supplies. The system will use solar energy for power. Since the container is essentially one closed module, researchers will not have to ask the astronauts to configure the servers.

    Why is there a supercomputer in space?

    Many of the calculations needed for research projects in space still occur on Earth due to the limited computing resources in orbit. Data transfer occurs with a delay, and if it is not so critical for projects in low Earth orbit, then with increasing distance from Earth and approaching Mars, it will become more and more. It will take about 20 minutes for the data to reach Earth, and it will take the same amount of time for the astronauts to get an answer. This will make research on the surface of the planet complex and potentially unsafe.

    Landing on the moon 50 years ago was at risk due to a computer error that occurred 8 minutes before landing. The on-board computer simply did not have the computing power to process all incoming data. Of course, modern smartphones, and indeed calculators or washing machines, already have significantly more powerful computing resources than those available to NASA in the 1960s. However, the needs have grown significantly.

    NASA’s Mars mission strategy requires astronauts to rely primarily on themselves. It is extremely important that computers can cope with unexpected input data and continue to miscalculate without transmitting them to Earth. To do this, you need a comprehensive data analysis that will cover all available sources and draw conclusions in real time, no matter where they come from: from cameras, sensors, navigation systems or a database with all the weather data ever collected on Mars.

    Data will come from a variety of sources. Sensors in wearable devices will constantly collect and process biometric information in order to track the slightest fluctuations in heart rate and other indicators. Cameras can analyze astronaut facial expressions to track aggression or stress levels. Navigation data will go to the on-board computer, which will need to constantly adapt the course, and in addition to monitor the condition of the equipment and the need for repairs. Powerful computing resources will be critical not only for automating routine tasks, creating simulations and working with artificial intelligence, but also for preventing disasters.

    What's next?

    After we see how Spaceborne Computer shows itself in space, in the next phases of the experiment, other computing systems, for example, Memory-Driven Computing (memory-oriented computing, see more in our article ) will go to the ISS . The benefit of Memory-Driven Computing for space projects and the mission to Mars is the ability to process data faster than the most powerful modern supercomputers. “We are creating an architecture that will be very flexible,” says Kirk Brezniker, chief architect of Hewlett Packard Labs. “No matter what new tasks come in, they will have computing power and enough memory to not only store data from each sensor, but also download data from previous missions.”

    Additional Resources and High Performance Computing News:

    1. HPE Supercomputer Portfolio Review: Record from the HPE Digitize Conference in Moscow
    2. High Performance Computing News: New Systems and Our Supercomputers in Top500 and Green500

    Also popular now: