1. The primer that repairs itself

Neal Stephenson’s The Diamond Age returns as the recalled work because the Young Lady’s Illustrated Primer is not a single device but a distributed system that must keep functioning despite damage, distance, and unreliable actors. The desktop is similar. Its computing fabric must keep working even when individual bits, registers, or cells are corrupted by radiation.

This entry applies the radiation literature to the desktop.

2. The desktop’s radiation-relevant features

The desktop has several characteristics that shape its radiation strategy:

  • Long operational life: it is meant to persist, so cumulative dose matters.
  • Modular cells: compute, storage, communications, and power functions are physically distributed.
  • COTS performance needs: high-throughput workloads will favor modern commercial processors and FPGAs.
  • Autonomous operation: ground intervention is not always available when a fault occurs.
  • Mixed criticality: some functions are mission-critical; others are customer payloads that can tolerate downtime.

These features push the desktop toward a hybrid, layered resilience strategy rather than a single rad-hard computer.

3. Derived requirements

The radiation literature gives the desktop six practical requirements:

  • Use rad-hard parts where destructive failure would end the mission. Power switches, critical watchdogs, and any component that cannot be protected by redundancy should be rad-hard or rad-tolerant.
  • Use COTS compute with system-level mitigation where performance dominates. EDAC-protected memory, TMR for critical logic, and periodic scrubbing can make COTS processors and FPGAs viable in LEO.
  • Distribute critical functions across cells. If one cell is reset by a latchup or SEFI, another cell should be able to hold platform-safe state until recovery.
  • Detect, report, and recover automatically. SEUs must be corrected or logged; SELs must be cleared by power cycling; uncorrectable errors must trigger rollback or safe mode.
  • Limit fault propagation. A radiation event in a customer payload cell should not corrupt attitude control or power management.
  • Characterize parts for the actual orbit and lifetime. Lot testing, proton-beam campaigns, and flight heritage data should inform the parts list rather than relying on generic COTS assumptions.

4. Precedents

Montana State University’s RadSat mission demonstrated a COTS FPGA-based computer architecture designed to mitigate radiation-induced faults through redundancy and reconfiguration. A recent MDPI paper on COTS embedded computers running cFS proton-tested a COTS single-board computer with flight software and confirmed that software-level fault tolerance can complement hardware-level mitigation.

These precedents support the hybrid approach: COTS performance where system-level mitigation is sufficient, rad-tolerant or rad-hard parts where it is not.

5. Interaction with earlier arcs

The radiation arc connects directly to:

  • the flight software arc: fault tolerance must be supported by architecture, message passing, and runtime reconfiguration;
  • the autonomous software update arc: updates must not corrupt radiation-protected state;
  • the software fault injection test arc: recovery procedures tested for software faults also apply to radiation-induced faults;
  • the predictive maintenance arc: error-rate trends are a leading indicator of radiation-induced degradation.

6. What this changes

  • The desktop inherits explicit radiation-resilience requirements.
  • The strategy is hybrid: rad-hard for mission-critical, non-redundant functions; COTS with EDAC, TMR, scrubbing, and distributed recovery for the rest.
  • The next entry will close the radiation reading arc and record the decision.