Monitoring and management: DDM, OSC, OCM, RFTS, alarms

Passive plant is silent; the active infrastructure around it is the only source of live information about a fibre link. Every transceiver measures its own light, every amplifier reports its pumps and gain, ROADM nodes watch the spectrum, supervisory channels carry it all between sites, and remote test systems fire OTDR pulses into the fibre while it carries traffic. This page sets out the layers of visibility, the alarms they raise and what each one means, how to keep baselines and trends, the protocols and tools that collect the data, how to correlate module data with fibre data to find the fault, and the practical rules for a monitoring setup that catches problems before customers do.

Layers of visibility

Layer	Source	What it sees	Detail
Transceiver diagnostics (DDM/DOM)	the module's monitor photodiodes and sensors, read by the host	Tx and Rx power, laser bias, temperature, supply voltage; on 400G+ and coherent modules also pre-FEC BER, chromatic dispersion, DGD, OSNR (VDM)	DDM basics, VDM & FEC metrics
Host port counters	MAC/PCS of the switch, router or NIC	link state and flaps, FEC corrected and uncorrected blocks, CRC errors, PCS errored blocks	Management & monitoring
Amplifier telemetry	EDFA and Raman units	input and output power, gain, pump current and temperature, loss of input, APR state	Amplifiers & regeneration
Optical channel monitor (OCM)	spectrum analyzer built into ROADM and terminal nodes	per-channel power and presence, OSNR estimate, tilt	Transport & access equipment
Optical supervisory channel (OSC)	a dedicated wavelength (1 510 or 1 610 nm) between line sites	span loss, remote alarms, management connectivity to unmanned sites	same
Remote fibre test system (RFTS)	OTDR units coupled at 1 625 or 1 650 nm to live fibres	new events, loss changes, break location on the fibre itself	Maintenance & restoration
Environmental sensors	cabinets, closures, huts	door, temperature, humidity, water, mains and battery state	same
Records and baselines	commissioning tests and the link passport	what "normal" is for every fibre and module	Documentation & labelling

Alarms and what they mean

Alarm	Raised by	Meaning	First check
Loss of signal (LOS)	receiver: no light	fibre cut, far-end transmitter off, wrong wavelength or channel, disconnected cord	far-end Tx power and link state (No link)
Loss of lock, loss of frame (LOL, LOF)	CDR, PCS, OTN framer	light present but unusable: speed or FEC mismatch, dispersion or noise beyond tolerance, very low power	Rx power against sensitivity, speed and FEC settings
Rx power low warning or alarm	module thresholds	plant loss grew or far Tx fell; typical thresholds a few dB above sensitivity	trend of far Tx and near Rx (Thresholds & alarms)
Rx power high	module thresholds	too little loss for ER/ZR or amplified output into a receiver	attenuator
Tx power low, laser bias high	module	laser ageing or failure	bias trend (Tx bias & laser ageing)
Temperature high	module or equipment	blocked airflow, failed fan, hot site, module above its class	site environment (Temperature & voltage)
Pre-FEC BER rising, uncorrectable blocks	host or module FEC statistics	the margin is being consumed — by power, dispersion, PMD, reflections or a dirty connector — before the link fails	correlate with Rx power and OSNR
Link flapping	host	intermittent contact, reflections, marginal power, thermal cycling	Link flapping
Module not present or unsupported	host	seating, coding, incompatible type	Vendor lock
Amplifier loss of input, APR active	amplifier	upstream fibre cut; pumps shut down for safety	do not defeat APR; locate the cut (Safety & handling)
Span loss increased (OSC)	line system	fibre degradation, bend, water, dirty connector in the span	RFTS or OTDR the span (Reading an OTDR trace)
RFTS new event or loss change	remote OTDR	a new bend, splice degradation, water or a cut at a located distance	convert to route position, dispatch
Door open, water, mains failed, battery low	environmental	the site is about to become a fibre cut	dispatch before the batteries end

Alarm hierarchies matter: a fibre cut raises LOS on every wavelength, loss of input on the amplifier, span loss on the OSC and a dozen service alarms — the root cause is the one lowest in the stack.

Baselines and trends

Practice	Rule
Link passport at commissioning	far Tx, near Rx, bias and temperature of both modules; OTDR both directions; loss at all wavelengths
Alert on deltas, not on absolute thresholds alone	2 dB below the baseline Rx is a call to look, long before the module's own low-power alarm
Poll interval	1–5 minutes for power values; counters per minute; RFTS scans hourly to daily
Seasonal behaviour	±1 dB swings on aerial and outdoor plant are normal — document them
Re-baseline after every change	new module, re-splice, re-patch
Accuracy	DDM is ±1–2 dB absolute but ±0.1 dB repeatable — perfect for trends, not for certification (Accuracy & limits)
Retention	keep at least a year of power history to see slow degradation and seasonality (Monitoring)

Protocols and tools

Mechanism	What it delivers	Notes
CLI (`show interface transceiver` and equivalents)	DDM values per port on demand	scripting-friendly; syntax per NOS (Reading DDM tools)
SNMP polling and traps	DDM, counters, equipment alarms	vendor MIBs for optical values; ENTITY-SENSOR-MIB on some platforms
Streaming telemetry (gNMI, OpenConfig, NETCONF)	DDM and counters every few seconds	`openconfig-platform-transceiver` models; best for trends
ethtool, mlxlink and NIC tools	DDM from server NICs	NIC tools & diagnostics
Line-system EMS/NMS	amplifier, OCM, OSC data and optical-layer alarms	TL1 or NETCONF northbound
RFTS software	trace comparison, event alarms, GIS mapping of distances	needs the route records to map distance to place
Umbrella NMS, time-series database and dashboards	correlation of fibre, module, service and environmental data	the place where root-cause analysis happens

Correlating module and fibre data

Observation	Likely cause	Confirm
Near Rx fell, far Tx and bias unchanged	plant loss increased: bend, connector, water, closure	OTDR or RFTS; inspect connectors
Far Tx fell, its bias rose	far-end laser ageing	replace module (Tx bias & laser ageing)
Both directions degrade together	common-path problem: a shared closure, duct, cabinet temperature	site check
Rx normal, pre-FEC BER rising	dispersion, PMD, reflections, non-linear or noise — not power	OSNR (coherent VDM), reflectance, characterization (Fibre characterization)
Rx swings daily by 1–2 dB	temperature on outdoor plant or a marginal connector	correlate with weather; inspect
Flaps with good average power	reflections, intermittent contact, dirty endface	reflectance on OTDR, endface inspection (Endface inspection & cleaning)
Temperature high with normal traffic	airflow or fan	site
Everything down at one site	power	environmental alarms, batteries
One wavelength down, others fine	that module, its patch cord, or a ROADM channel setting	swap module, check channel plan

Practical rules

Rule	Why
Monitor both ends of every link	plant loss is far Tx minus near Rx; one end tells you nothing about which side degraded
Read DDM through the host, and keep the bench value	the host's reading is the operational truth; the bench value is the baseline before the plant (DDM in the app)
Alarm thresholds per link, not per module type	the module's factory thresholds are wide; the link's margin is specific
Test the protection paths	a standby path with no traffic still needs its Rx power watched, or it fails silently
Monitor the monitors	RFTS units, OSC terminals and sensors fail too
Keep the records current	an alarm at "12.3 km" is useless without a route map (Documentation & labelling)
Separate the management network	an outage must not take down the ability to see it

In CodingBox

The bench is the first monitoring point: CodingBox reads a module's DDM, thresholds and VDM capabilities before installation, so the link passport starts with known-good values for the module alone — Tx power, bias, temperature — and the host's later readings can be judged against them (DDM in the app, Check transceiver, Monitoring).