Configuration

This guide covers the configuration options for ai_nn_controller.

Key Insight: Registration-Based Discovery

Nodes and AIC apps only need to know the register’s address. The register returns the broker’s address during the registration handshake. This means you never need to configure broker addresses in nodes or applications — only the single register address.

Configuration Files

Register Configuration

The register service (controller_components/register/register.conf) defines the central configuration:

# Register's own binding address
registration_ip_address = aic_register
ip_address = 0.0.0.0

# Broker address (returned to registrants)
databus_ip_address = node_msg_broker

# Port mappings
recv_measurements = 5555    # Nodes PUSH measurements here
pub_measurements = 5554     # Apps SUB to measurements here
recv_commands = 5556        # Apps PUSH commands here
pub_commands = 5557         # Nodes SUB to commands here
register_sub = 5558         # Registration port (REQ/REP)

AIC App Configuration

Apps only need to know the register address (aic_app.conf):

ip_address = aic_register
register_port = 5558

Node Configuration

Nodes using the controlled_entity package need a node.conf file with the register address and a unique node ID. The NodeRunner parses this file and injects the values into self.config on the node instance.

Required keys:

# Register address (Docker service name or IP)
ip_address = aic_register

# Registration port
register_port = 5558

# Unique node identifier — must not conflict with other nodes
node_id = 3

Commonly-set optional key:

# Port this node publishes measurements on (varies per node)
pub_port = 5562

Optional / custom keys — add any key-value pairs your node needs:

# For srsRAN nodes (InfluxDB connection)
influxdb_url = http://influxdb:8086
influxdb_token = <your-token>
influxdb_org = srs
influxdb_bucket = srsran
poll_interval = 1

# For Twilight optical nodes
twilight_url = http://host.docker.internal:8000
device_name = Amp_ila1
device_type = amplifier

Custom keys are accessible in the node’s setup() and poll_measurements() methods via self.config:

def setup(self):
    url = self.config.get("influxdb_url", "http://localhost:8086")

The parser auto-converts values to int or float where possible. Lines starting with # are comments.

Port Summary

Port Allocations
Port	Service	Socket Type	Direction	Purpose
5554	Broker	PUB	→ Apps	Measurement publication
5555	Broker	PULL	← Nodes	Measurement collection
5556	Broker	PULL	← Apps	Command collection
5557	Broker	PUB	→ Nodes	Command publication
5558	Register	REP	↔ All	Registration (REQ/REP)
6379	Redis	TCP	↔ Register	State storage
8000	AIC Server	HTTP	↔ Clients	REST API & MCP

AicController Configuration

When instantiating AicController, you can configure:

from ai_nn_controller.AicController import AicController

controller = AicController(
    with_api=True,       # Enable FastAPI server
    api_host="0.0.0.0",  # API host address
    api_port=8000,       # API port
    verbose=True         # Enable verbose logging
)
controller.run()

Command-Line Arguments

When running an app directly:

python aic_app.py --verbose --port 8000 --host 0.0.0.0

Arguments:

--verbose, -v: Enable verbose output for debugging
--port, -p: Port for the FastAPI server (default: 8000)
--host: Host address for the FastAPI server (default: 0.0.0.0)

Environment Variables

The following environment variables are recognised at runtime:

Environment Variables
Variable	Default	Effect
`AIC_MAX_QUEUE_SIZE`	`500`	Per-app per-topic measurement queue depth (controller side)
`AIC_ARBITRATION_STRATEGY`	`last_write_wins`	Arbitrator strategy: `last_write_wins` or `min_gap`
`NODE_MAX_PENDING_COMMANDS`	`200`	Max pending commands + idempotency cache size on node side

Application Configuration

Within your application, configure via class attributes:

@aic_app(name="MyApp")
class MyApp(AicApp):
    aic_app_id = 100

    # How often process() is called (seconds)
    control_loop_update_time = 2

    # Measurements to read per node
    read_measurements = {
        3: ["gain", "power", "temperature"],
        4: ["status", "alarm"],
        8: ["preamp_gain", "booster_gain"]
    }

    # Control functions per node
    control_functions = {
        3: ["SET_GAIN"],
        8: ["SET_GAIN", "SET_VOA", "SET_TILT"]
    }

    # cell_ids is auto-generated as the union of
    # read_measurements and control_functions keys: [3, 4, 8]

Next Steps

Developing Control Applications — Create custom control applications
Developing Network Nodes — Build network nodes with controlled_entity
Docker Deployment — Docker deployment guide
REST API Reference — REST API configuration