Over the past two months, Hewlett Packard Enterprise (HPE), Windhover Labs, and GSI Technology have announced new computing resources for space applications. HPE has a second-generation server now onboard the International Space Station, while Windhover Labs and GSI have parts and silicon targeted towards satellite applications.
The HPE Spaceborne Computer-2 (SPC-2) was sent to the International Space Station (ISS) onboard the February 20, 2021 Northrup Grumman NG-15 supply mission. It is an improved version of the first Spaceborne Computer developed and sent ISS in 2017 for a one year test. HPE wanted to demonstrate that a commercial off-the-shelf server using software-based hardening/fault tolerant features could survive launch and successfully operate in the spartan, zero-g and higher radiation environment of the space station.
In-space computing resources onboard ISS enable astronauts and researchers to process on-site/at the edge, delivering faster results rather than having to download data through relatively limited ISS bandwidth and process it on the ground. As explorers push out farther away from Earth onto the Moon and Mars, bandwidth will become more precious and communication delays will increase, requiring more powerful local computing capabilities for many tasks.
Built around the HPE Edgeline Converged EL4000 Edge System and HPE DL360 ProLiant server, HP SPC-2 will deliver twice the performance of the original Spaceborne Computer and is equipped with NVIDIA GPUs to support faster imaging processing and will also support specific AI and machine learning products. The new computer will enable real-time monitoring of astronaut health by processing X-ray, sonograms and other medical data quickly and analyzing imaging, signal and other data directly.
SPC-2 is expected to support onboard research for the next two to three years. It is also connected to the Microsoft Azure cloud, enabling researchers to seamlessly pass off larger compute problems directly to a scalable computing solution.
Edge computing start-up OrbitsEdge plans to incorporate HPE Edgeline systems into its SatFrame solution for satellite usage, providing an on-demand in-orbit computing capability for users both on the ground and in-space.
Windhover Labs announced its low-cost modular flight computer for commercial drones and small satellites on March 11, 2021. The OBC computer is the size of a pack of playing cards and is the company’s first commercial product for the small satellite industry. It comes with a quad-core ARM Cortex-A53, a dual-core ARM Cortex-R5,2 GB of RAM, 128MB of flash, a FPGAs, options for adding a GPU and a VCU for various applications and uses the CFS open source flight software system. Co-founder Mathew Benson said the company is in the process of filling out its website and anticipates the computer to be in demand for both high-end drone and spacecraft application.
This week, GSI Technology and Space Micro announced they had been awarded a NASA contract to develop a real-time data sorting interface processor (IPU) to boost processing power onboard Earth observation missions. The GSI/Space Micro IPU plugs into the main communications bus to add enhanced in-line processing. Designed for AI/ML applications, GSI’s Gemini Associative Processing Unit (API) complements its radiation-hardened and space-qualified RAM, enabling in-line parallel processing and computing directly within a memory array. For faster performance and redundancy, multiple processors can be included on the same board and you can stack multiple units together.
Making more compute power available for imaging satellites can provide quicker products that can be directly downloaded to users, rather than having to download an unprocessed image into the cloud, clean it up for use, then deliver it to the customer. Synthetic aperture radar (SAR) Capella Space uses a NVIDIA processor onboard its satellites to do some quick imaging processing so uses can get a rough look at an image while a larger data file awaits downloading and processing on the ground.
