Storage Management Compute Host
Rad-Hard SBC Hosting MGMT Software Stack
Dedicated rad-hard single-board computer hosting the L2-ESS-MGMT software stack (SOC fusion, SOH prognostics, dispatch optimizer, charge controller, safety monitor) — provides isolated, deterministic compute for the storage system's logic.
Purpose
Provide a single physical platform that runs all storage-management software with guaranteed CPU/memory budgets and hardware fault containment, independent of the central L1-CDH compute pool to limit blast radius of CDH issues.
Context
Houses all L3-ESS-MGMT-* software; talks to L1-CDH for high-level commands, time, and persistent storage; talks via internal high-rate bus to all storage controllers. Implements bus interface to L3-CDH-BUS for inter-system communication.
Principles
- ▸Dedicated single-purpose SBCs improve fault containment by limiting blast radius
- ▸Rad-hard SoCs (BAE RAD5545, GR740, Vorago VA10800) provide redundancy and high MTBF
- ▸Real-Time Operating Systems (RTEMS, VxWorks, LynxOS) provide deterministic scheduling for control loops
- ▸Memory protection unit (MPU) prevents one task from corrupting another
- ▸Hardware watchdog timer recovers from software hangs
- ▸Triple-buffered storage of critical parameters survives SEU corruption
Typical implementations
- ▸BAE RAD5545 PowerPC SoC (rad-hard quad-core)
- ▸Cobham GR740 LEON4FT SPARC (ESA flagship)
- ▸Vorago VA10800 ARM Cortex-M0 (rad-hard, lower power)
- ▸Microchip PolarFire SoC FPGA with embedded RISC-V cores
- ▸ISS battery controller compute hosts heritage
- ▸NASA cFS framework executing on each of these platforms
Lunar considerations
- ▸100-year operation requires platform replaceable as architecture evolves; standardized bus interfaces aid migration
- ▸Rad-hard SBCs are scarce and expensive — design constraints favor lean software stack
- ▸Watchdog and reset mechanism must survive SEU on watchdog logic itself
- ▸Reduced-power 'survival mode' when reactor power drops; controller still must operate at minimum CPU clock
- ▸Field firmware update via authenticated DTN bundles (see L3-COM-CRYPT-AUTH); supports staged update with rollback
Specifications
Functional
| primary function | Provide rad-hard compute platform for storage management software |
| inputs | Sensor inputs from all L3-ESS-* controllers, Commands from L1-CDH and L2-CDH-AUTO, Time from L3-COM-TIME-DIST, Power from L1-PDM |
| outputs | Software outputs from each MGMT subnode, Telemetry to L1-CDH, Health/heartbeat to L1-CDH safety supervisor |
| cpu core count | 4 |
| cpu clock mhz | 200 |
| memory ram mb | 512 |
| non volatile storage gb | 8 |
| watchdog period s | 0.5 |
| supported rtos | RTEMS, VxWorks-178C, Linux RT |
| radiation tid krad | 100 |
| seu immune | True |
Physical
| mass kg | 1.5 |
| dimensions | 180 × 120 × 30 mm 3U cPCI board |
| materials | RAD5545 or GR740 SoC die, Rad-hard SRAM with EDAC, Rad-hard NOR flash for code, MRAM for critical state, Polyimide PCB, Aluminum chassis with EMC gasketing |
| operating temperature c | -40, 70 |
| radiation tid krad | 100 |
| seu let threshold mev cm2 per mg | 80 |
Operational
| power consumption w | 12 |
| thermal range c | -40, 70 |
| lifetime years | 25 |
| mtbf hours | 300000 |
Interfaces
Provides
- Compute platform for all MGMT software
- Telemetry, status, and heartbeat
Requires
- 28 V regulated DC, ~12 W
- Cold-plate interface
- Time reference for control loops
- Inter-subsystem data bus access
Cite this entry
Lunar Ark Codex. "Storage Management Compute Host" (L3-ESS-MGMT-PROC). Retrieved 10 September 2026, from https://lunarark.com/entry/L3-ESS-MGMT-PROC
Licensed CC-BY-SA 4.0. You may reuse and adapt this entry with attribution, under the same licence.