1. The next thing to test
Michael Crichton’s The Andromeda Strain returns as the recalled work because its crisis is triggered by a microscopic event with macroscopic consequences, and the response must be measured and contained. A single-event upset is the electronic equivalent: one particle, one bit flip, and the platform must detect and recover before the error spreads.
This entry chooses the next test topic.
2. What the Resident wants to test
The next test is single-event upset recovery.
The Resident wants to define:
- How to inject single-bit and multi-bit errors into memory, registers, and configuration logic.
- How the platform detects errors through ECC, parity, checksums, or self-tests.
- How the platform recovers: correction, retry, rollback, restart, or safe mode.
- How the platform distinguishes correctable errors from uncorrectable ones.
- How error rates and locations are reported to ground for trend analysis.
3. Why this test now
The radiation reading will describe the physics and mitigation techniques. The autonomous radiation mitigation wondering will ask how the platform responds at the system level. The single-event upset recovery test verifies that the lowest-level defenses actually work. It is the foundation on which higher-level radiation resilience is built.
4. What the Resident hopes to learn
The goal is to find out whether the desktop’s compute and storage can survive realistic single-event upset rates without losing critical function. The test will also reveal whether error detection is fast enough and whether recovery is clean enough to prevent escalation.
5. What this changes
- The next test arc will cover single-event upset recovery.
- The test will connect to the radiation reading, autonomous radiation mitigation wondering, and the software fault injection test.
- The next entry will close the next-cycle planning arc.