Linux in Space: Fault Detection, Recovery and Fault-Tolerant System Designs
summary
Linux runs as a payload controller on CubeSats using commodity SBCs (Raspberry Pi, BeagleBone) with Yocto-built nominal and recovery images, U-Boot boot slots and a recovery flag → recovery image flow; memory uses ECC and duplicated N‑flash or dual SD strategies, sometimes full CPU duplication and voting. Real missions: Visat has two Linux nodes working >2.5 years, thermal vacuum cycling ~1,000 hours with CPU peaks ~80°C, radiation-tested components, heavy vibration testing. Telemetry is slow (9.6 kbps, initial packet loss >50%; future downlink ~10 Mbps for ~20 min/day). Launch can be ~€100k (~10% of project).
AI-written summary · about
chapters (18)
- 0:00Linux in space projects
- 3:09Linux in space challenges
- 6:05Fault tolerance fundamentals
- 9:17Redundancy and fail-safe design
- 13:11Software updates and boot reliability
- 16:41Real-world satellite Linux systems
- 18:46Linux computer for CubeSats
- 21:40Mission-level fault mitigation
- 22:47Avoiding single points of failure
- 26:23System-level fault mitigation
- 27:06Hardware testing for space reliability
- 28:10Thermal vacuum testing
- 31:07Simplicity in autonomous design
- 33:20Minimizing writes to NVM
- 34:02Yocto-based Linux build
- 35:07Boot procedure and recovery
- 37:38Yocto layers for space missions
- 38:42Q&A on speeds and costs
related talks
38mLinux for Space - Mission Begins
47mBoldly Going, Running Linux in Space
42mVZLUSAT-2: CubeSat with a Linux Payload Computer
32mUsing Linux as an Embedded OS for the Space Industry
46mLinux on Quick Turnaround Projects at Ball: No, We Aren’t Putting Linux in Canning Jars
16mTuning Linux for Embedded Space Applications
16mKeynote: Space Grade Linux: Building a Safer, Open Source Future for Space Systems
49mHARDWARE in SPACE! The Kernel at the Edge of the Universe