1. The ship that diverts power to the hull

James S. A. Corey’s The Expanse returns as the recalled work because the Rocinante is constantly routing power, water, and air to wherever the current crisis demands. A radiation event is a similar crisis: the platform must decide, in seconds or minutes, where to put its limited resilience budget.

This entry wonders about throttling and relocating work during radiation events.

2. Throttling

Throttling means reducing activity to lower the probability that a radiation-induced error becomes a system failure. Options include:

  • Reduce clock speed so that transient errors are less likely to propagate before being caught.
  • Pause non-critical tasks so that only essential functions are running.
  • Batch or defer writes to storage and flash memory.
  • Disable speculative execution or other performance features that increase vulnerability.
  • Slow down communication bursts to reduce the chance of a corrupted command or telemetry frame.

Throttling trades performance for safety. The platform needs a policy that defines how much performance can be sacrificed and for how long.

3. Relocation

Relocation means moving active work to a different part of the platform. Options include:

  • Move tasks to a better-shielded cell. A cell deeper inside the structure, behind battery mass or thermal hardware, sees lower flux.
  • Move tasks to a rad-tolerant or rad-hard cell. Some cells may use rad-hard processors specifically for emergencies.
  • Move storage operations to a replicated volume. If one storage cell is seeing high error rates, write to a different one.
  • Move decision authority to a simpler, more robust controller. A small rad-hard microcontroller may take over if the main compute complex is uncertain.

Relocation requires the desktop’s software architecture to treat cells as relocatable in the first place, which the flight software reading arc already identified as a goal.

4. Coordinating throttling and relocation

Throttling and relocation should not be independent. A high-flux event might trigger:

  1. immediate throttling of non-critical payloads;
  2. migration of critical tasks to the best-shielded cell;
  3. increased scrubbing on all active memory;
  4. notification to ground with the current risk level;
  5. return to normal operations when flux drops and error rates normalize.

5. What limits this

  • The platform may not have enough spare cells to relocate everything.
  • Relocation takes time and bandwidth; during a sudden event it may not complete before the peak.
  • Not all workloads can be paused or moved. Attitude control, thermal management, and fault protection must run continuously.
  • Overlapping events, such as a solar particle event during an SAA pass, may exhaust mitigation capacity.

6. What this changes

  • Throttling and relocation are practical autonomous responses if the architecture supports cell migration and policy-driven performance limits.
  • The desktop should design its compute and storage layout with radiation-safe havens in mind.
  • The next entry will ask what autonomous radiation mitigation needs from the desktop architecture.