1. The universe does not cooperate

Liu Cixin’s The Three-Body Problem begins with the discovery that the laws of physics are not as stable as humanity assumed. The universe is stranger and more hostile than the characters want it to be. The desktop faces a smaller version of the same realization: LEO is not a clean room. It is a place where protons, electrons, and heavy ions pass through the hull and flip bits without asking permission.

This entry synthesizes what the radiation literature means for the desktop.

2. The desktop is a computer that consumes more than 500 watts

That identity drives the radiation question. A high-power compute platform needs dense processors, dense memory, and dense storage. Density works against radiation tolerance: smaller nodes and smaller cells are more sensitive to single-event effects, even if they are sometimes more tolerant to total ionizing dose. The desktop cannot simply choose the most rad-hard parts if those parts cannot deliver the needed performance per watt.

3. What the literature says the desktop should do

The reading points to a layered strategy:

  • Select COTS parts with known LEO heritage or test data. The MDPI survey and NASA reports make clear that many modern COTS devices survive LEO dose levels, but the variance is large. A part that works in one mission may not work in another orbit or batch.
  • Design for SEE tolerance at the system level. This means error-correcting memory, scrubbing, watchdog timers, redundant compute nodes, and the ability to reboot or failover a cell without losing the platform.
  • Protect storage with end-to-end correction. LDPC, RAID parity, and periodic scrubbing turn a radiation-sensitive flash device into a radiation-tolerant storage system.
  • Measure the actual environment. Model predictions vary by factors of 3 to 30. An on-orbit dosimeter or radiation monitor gives ground truth and lets the operator adjust operating margins or retire a cell before it degrades.
  • Plan for graceful degradation. A cell that starts at nominal performance may drift. The desktop should be able to identify a degrading cell, migrate its workload, and replace it.

4. What the literature does not answer

The literature does not yet give a simple shopping list for a 500-watt COTS compute platform in LEO. The exact SEU rate depends on orbit, shielding, device cross-section, and solar cycle. The exact TID depends on the same variables plus the mission duration. The exact trade-off between performance and radiation tolerance depends on the specific processors and storage controllers chosen.

What the literature does provide is a framework: test the parts, correct the errors, monitor the environment, and accept that faults will happen.

5. What this changes

  • The desktop’s compute and storage designs must include explicit radiation-mitigation architecture, not just hope for tolerant parts.
  • Error correction and failover are as important as the raw compute specification.
  • On-orbit radiation monitoring is added to the list of capabilities the platform should carry.
  • The next entry will close the radiation reading arc and decide what to do with this knowledge.