Many IP and data-center engineers eventually find themselves looking at optical alarms: a DWDM link between data centers, a leased wavelength or a transponder shelf in the corner of the meet-me room. Optical terminology can feel like a different language. This guide translates the most important alarms and measurements into terms any network engineer can use.
A quick model of an optical link
A typical DWDM service between two sites looks like this:
- A client signal (for example 100GE from a router) enters a transponder or muxponder.
- The transponder maps it into an OTN (Optical Transport Network) frame and transmits it on a specific wavelength.
- Multiplexers combine many wavelengths onto one fiber.
- Amplifiers boost the signal along the route, often every 60–100 km.
- At the far end, the process reverses.
Alarms can appear at every layer: the fiber and optics (physical), the OTN framing (digital) and the client signal.
The core defect alarms
LOS: Loss of Signal
The receiver sees no light, or light below its sensitivity threshold. This is the optical equivalent of "cable unplugged".
Likely causes: fiber cut, disconnected patch cord, failed transmitter at the far end, or an amplifier failure upstream.
Tip: LOS on many wavelengths at once points to a line or fiber problem (cut, amplifier). LOS on a single wavelength points to that channel's transponder or patching.
LOF: Loss of Frame
Light is present, but the receiver can't find the frame structure. The signal is too degraded to decode, or there is a configuration mismatch such as the wrong rate or FEC mode.
AIS: Alarm Indication Signal
A downstream indication that something upstream has failed. When a node detects a failure, it sends AIS forward so downstream equipment knows the problem is elsewhere and doesn't raise its own root-cause alarms.
Why it matters for correlation: AIS is a symptom. Your fault management system should treat AIS as secondary to the LOS or LOF that caused it.
BDI / RDI: Backward Defect Indication
The far end tells the near end, "I'm not receiving you properly." In OTN this is BDI; in SONET/SDH terminology it is RDI (Remote Defect Indication). If you see BDI on your side, the problem is in your transmit direction.
Signal degrade and signal fail
Threshold-based alarms. The link still works, but the error rate has crossed a warning level (degrade) or a critical level (fail). These are your early warnings before a hard outage.
Performance measurements that matter
Optical power (dBm)
Transmit (Tx) and receive (Rx) power in dBm. What matters is whether Rx power sits within the receiver's specified range, with margin.
- Too low: dirty connectors, a bad splice, a bent fiber or excessive loss
- Too high: can overload the receiver, which is common on short links without attenuation
Track Rx power trends over time. A slow decline often means a degrading connector or fiber before it becomes an outage.
Span loss
The loss between two amplifiers or sites, in dB. Standard single-mode fiber loses roughly 0.2–0.25 dB per km at 1550 nm, plus connector and splice losses. A sudden increase in span loss is a strong sign of physical damage or a bad patch.
Pre-FEC BER
Modern coherent optics use Forward Error Correction (FEC) to fix bit errors. Pre-FEC BER is the error rate before correction, and it is the best early-warning metric on an optical link. Post-FEC BER should normally be zero. If post-FEC errors appear, FEC is being overwhelmed and traffic is being affected.
Monitoring tip: alert on rising pre-FEC BER trends, not just on post-FEC errors. By the time post-FEC errors appear, customers are already affected.
OSNR and Q-factor
Optical Signal-to-Noise Ratio measures how clean the signal is relative to the noise added by amplifiers. Q-factor is a related quality measure derived from signal statistics. Both decline as a link degrades, and both have minimum values that depend on the modulation format and FEC.
Chromatic dispersion and PMD
Physical effects that spread pulses over distance. Coherent optics compensate for them digitally, but the reported values help diagnose problems on very long or older fiber.
Performance monitoring bins: ES, SES and UAS
Transport equipment records performance in standard time bins, typically 15-minute and 24-hour intervals, with counters such as:
- ES (Errored Seconds): seconds with at least one error
- SES (Severely Errored Seconds): seconds with a high error rate
- UAS (Unavailable Seconds): seconds the path was considered unavailable, which is key for SLA reporting
These bins are ideal for trend reports and SLA reporting. Collect them into your performance platform so optical health sits alongside IP metrics.
The OTN layers in one paragraph
OTN wraps client signals in layers: the OPU carries the payload, the ODU adds path monitoring end to end, and the OTU adds section monitoring and FEC between optical regenerators or transponders. Alarms are reported per layer (for example OTU-LOF or ODU-AIS), which tells you where in the path the problem sits. Tandem Connection Monitoring (TCM) adds monitoring points so carriers can isolate faults at hand-off boundaries between their network and another operator's.
Correlating optical and IP alarms
When a DWDM link fails, you usually see:
- Optical: LOS on the line side, AIS downstream, BDI back toward the source
- Client: link down on the router interfaces using that wavelength
- IP: routing adjacency loss, traffic shifting to alternate paths
Good correlation shows the optical LOS as the root cause and the router interface and routing alarms as symptoms. That requires your fault management system to know which router ports ride on which wavelengths, another reason accurate inventory and service mapping matter.
Key takeaways
- LOS and LOF are root causes. AIS and BDI tell you where the problem is relative to you.
- Pre-FEC BER, Rx power and OSNR trends are the best early warnings.
- Collect 15-minute and 24-hour PM bins (ES, SES, UAS) for trends and SLAs.
- Map router ports to optical services so optical failures correlate with the IP impact they cause.