- Dedicated Watchdogs: Employ a custom-designed watchdog monitor (WDM) for each microprocessor to detect potential malfunctions.
- WDM Reliability: Ensure WDMs are independent (with separate power and clock sources) and undergo regular checks to confirm their operational status.
- Tailored Design: WDMs should align precisely with the specifications and critical functions of the microprocessor they are monitoring.
- Failure Reporting: WDMs need robust failure signaling to alert other system components, enabling appropriate corrective action.
Software Monitor Considerations
- Additional System and Box-Level Built-in-Test (BIT):- Start-up BIT: Executed at system power-on.- Interface, Initiated, Interruptive, or Intermittent BIT (IBIT): As needed.- Periodic BIT (PBIT): For regular checks.- Maintenance BIT (MBIT): For in-depth diagnostics.
- Additional Voting Planes: Utilize multiple software layers or processes to compare results and flag discrepancies, enhancing reliability.
- Data Integrity Checks: Verify the accuracy and consistency of data.
- Configuration Register Monitors: Track critical microprocessor settings to identify unauthorized or unexpected changes.
Hardware Monitor Considerations
- Memory Monitors:- Parity/Error Correction Code (ECC): Detects and corrects memory errors for improved reliability.- SEU (Single Event Upset) Monitors: Identifies radiation-induced memory changes.
- Bus Monitors (Internal & External):- Address and data parity checks: Maintain data integrity on buses.- Protocol checks: Validate adherence to correct communication protocols.- Checksums and Cyclic Redundancy Code (CRC): Error detection techniques.- Activity pattern monitoring: Identifies unusual bus behavior.
- Discrete Digital Signals Monitoring:- Dualize I/O: Create redundant input/output for comparison purposes.- BIT (Built-in-Test) stimulation capability.- Wraparounds: Test signal paths for closed-loop integrity.
- General Signal Monitoring:- Wraparounds: Test signal paths (both analog and digital).- Range checks: Ensure signals remain within expected values.- Rate checks: Monitor the timing of signals.
- Dissimilar Hardware: Introduce redundancy and reduce common-cause failures by using diverse hardware types.
- Configuration Register Monitors: Track critical processor settings.
