EPON Manager – Component Design

Created on March 13, 2026

1. EPON Controller (Main Component)

Responsibilities

  • Initialize all subsystems (logger, RBus, telemetry, HAL)
  • Coordinate between threads and modules
  • Main event loop and lifecycle management
  • Resource cleanup on shutdown

Initialization Sequence

epon_controller_init()
  > logger_init()
  > rbus_thread_start()
  > telemetry_init()
  > epon_data_init()
  > hal_event_listener_start()
  > hal_init()
      - Register onu_status_callback
      - Register interface_status_callback
      - Register alarm_callback
  > stats_polling_thread_start()

Key Functions

  • System initialization and startup
  • Thread creation and management
  • Graceful shutdown handling
  • Error recovery coordination

2. HAL Event Listener Thread

Responsibilities

  • Receive events from EPON HAL via registered callbacks
  • Parse and categorize incoming events (ONU status and interface status)
  • Dispatch events to appropriate handlers
  • Maintain event processing statistics

Event Types Handled

ONU Status Callback Events (epon_onu_status_t)

Simplified overall ONU status. 

These events are for internal state tracking and logging only—they do NOT trigger WanManager updates.

Respective epon_onu_interface_status_t should be updated by the HAL in case of ONU status changes.

  • EPON_ONU_STATUS_LOS (Loss of Signal):
    • Update internal EPON ONU status to DOWN
    • Log event with WARNING severity
    • Raise telemetry event
  • EPON_ONU_STATUS_DOWNSTREAM_SIGNAL_DETECTED:
    • Update internal EPON ONU status to Initializing
    • Log transition state
  • EPON_ONU_STATUS_REGISTRATION:
    • Update internal EPON ONU status to UP
    • Log successful registration
    • Raise telemetry event
  • EPON_ONU_STATUS_DEREGISTRATION:
    • Update internal EPON ONU status to DOWN
    • Log deregistration event
    • Raise telemetry event

Interface Status Callback Events (epon_onu_interface_status_t)

PRIMARY EVENT: Per-interface status changes for virtual interfaces (veip0, veip1, etc.). These events trigger WanManager updates.

  • Interface LINK_UP (e.g., veip0):
    • Update interface status in internal cache
    • Add interface to active interface list
    • Notify WanManager of interface availability via RBus (primary action) – If any of the interfaces go UP, WanManager should be notified that EPON PHY is UP – Update WanManager virtual interface list with interface name and status
    • Raise telemetry event for interface UP
  • Interface LINK_DOWN (e.g., veip0):
    • Update interface status in internal cache
    • Remove interface from active interface list
    • Notify WanManager of interface down via RBus (primary action) – If all of the interfaces go DOWN, WanManager should be notified that EPON PHY is DOWN – Update WanManager virtual interface list with interface name and status
    • Raise telemetry event for interface DOWN

Event Processing Flow

3. Stats Polling Thread (Harvester)

Responsibilities

  • Periodically poll statistics from EPON HAL
  • Update internal cache with fresh data
  • Push statistics to telemetry system
  • Maintain polling schedule

Configuration

  • Default Interval: 15 minutes (900 seconds)
  • Configurable: Via rbus
  • Optional: Can be disabled if not needed

Operation Flow

 Statistics Collected

  • Interface statistics (TX/RX bytes, packets, errors)
  • ONU operational parameters
  • Link quality metrics
  • Performance counters

4. RBus/DBus Thread

Responsibilities

  • Initialize and maintain bus connection
  • Register TR-181 DML parameters
  • Handle incoming GET/SET requests
  • Publish events to other RDK components
  • Update WanManager with PHY status and interface list

TR-181 Request Handling

Key Operations

  • DML parameter registration
  • Request validation and processing
  • Cache-aware GET operations
  • Event publication
  • WanManager interface

5. Telemetry Module

Responsibilities

  • Register with T2 telemetry system
  • Format and report telemetry events
  • Submit periodic statistics
  • Handle telemetry markers

Integration Points

  • T2 telemetry service
  • Event formatting and submission
  • Statistics aggregation
  • Marker management

Event Types

  • Link events: UP/DOWN transitions
  • Alarm events: Critical/Error alarms
  • Statistics events: Periodic metrics
  • Error events: System errors

6. Logger Module

Responsibilities

  • Initialize RDK logger subsystem
  • Provide logging API wrapper
  • Support multiple log levels
  • Optional HAL logging integration

Log Levels

LevelUsageExample
FATALSystem-critical errorsFailed initialization
ERRORComponent errorsHAL call failure
WARNINGAbnormal conditionsCache miss threshold
INFONormal operationsLink status change
DEBUGDetailed debuggingEvent queue details

Features

  • Runtime log level configuration
  • Component-specific logging
  • HAL library integration (optional)
  • Thread-safe logging

7. HAL Wrapper with Cache

Responsibilities

  • Provide abstraction layer for HAL calls
  • Cache statistics with TTL (Time To Live)
  • Check cache validity before making HAL calls
  • Thread-safe access with automatic expiration
  • Forward calls to HAL only when cache miss occurs
  • Memory management for cached data

Architecture


 Cache Operations

  • Get: Retrieve value with cache validity check
  • Set: Store value and invalidate if needed
  • Invalidate: Force cache expiration
  • Cleanup: Remove expired entries

Cache Policy

  • TTL: 30 seconds (default, configurable)
  • Strategy: Time-based expiration with lazy cleanup
  • Thread Safety: Mutex-protected (RW lock recommended)
  • Bypass: Controller can bypass cache for direct HAL calls

Error Handling

Strategy

  • Initialization Errors: Retry with backoff, log and exit if persistent
  • Runtime Errors: Log, attempt recovery, continue operation
  • Critical Errors: Notify watchdog, trigger restart if needed

Recovery Mechanisms

  1. HAL Failure: Retry with exponential backoff (max 3 attempts)
  2. RBus Disconnection: Auto-reconnect with retry logic
  3. Thread Crash: Watchdog detection and restart
  4. Cache Corruption: Clear and rebuild cache
  5. Memory Issues: Cleanup and resource management
Previous
EPON HAL Proposal
Next
EPON Manager – Configuration & Deployment
Go To Top