Choosing WDM for a Fiber Link: DWDM vs CWDM in Practice

The capacity problem you actually face

You have one fiber pair in the ground and a demand for 3 to 8 times more capacity. Trenching a new route is not an option. Wavelength Division Multiplexing lets you put multiple light colors on the same fiber. The useful takeaway is that WDM multiplies capacity, but the engineering decision is not just more wavelengths. It is choosing the right spacing and hardware for distance, cost and operational risk.

What WDM buys you and what it costs

WDM allows multiple wavelengths to coexist on a single fiber. In practice that means 10 to 100 times more capacity than a single wavelength system without new fiber.

Dense WDM uses tightly spaced channels about 0.8 nm apart in the C-band. That spacing supports 100 Gbps per channel and scales toward 400 Gbps with coherent modulation. It requires tunable lasers, arrayed waveguide gratings or thin-film filters for multiplexing, and erbium doped fiber amplifiers to hold signal power across wavelengths.

Coarse WDM uses wider spacing of 8 to 16 nm. It is lower cost, works with fixed-wavelength transmitters, and is typically limited to shorter reach. CWDM is common for campus and metro links under about 80 km where amplifier count is low.

Both variants share the same physical impairments. Chromatic dispersion spreads pulses with wavelength. Polarization mode dispersion causes random delay between polarization states. Inter-channel crosstalk grows with channel count and filter selectivity. These effects increase with distance and channel density.

Worked example: 80 km metro with three services

Scenario: two sites 80 km apart, one fiber pair available, three services needing 100 Gbps each.

Option A DWDM: 100 GHz grid, three channels at 1549.32 nm, 1550.12 nm, 1550.92 nm. Requires DWDM mux/demux, EDFA at each end, dispersion compensation modules. Higher capex for tunable lasers and AWGs, but reach and future expansion to 40+ channels is preserved.

Option B CWDM: channels at 1530 nm, 1550 nm, 1570 nm. Fixed lasers, passive mux/demux, no amplifiers if loss budget allows. Lower hardware cost. Risk is that 80 km may be near the limit for 100 Gbps over CWDM without amplification, and future channel additions are constrained by 18 channel max and wider spacing.

Decision rule used here: if the link will need more than 8 wavelengths, or amplifiers are already planned, DWDM is the conservative choice. If the link stays under 80 km, three wavelengths, and budget is constrained, CWDM can be acceptable with margin on loss.

Verification before you sign off

Wavelength alignment and power balance are the first failure modes.

• Use an optical spectrum analyzer at the mux output to confirm channel spacing and power levels. Each wavelength should fall within its designated band and within 1 to 2 dB of the target power.
• Perform bit error rate tests on each channel after amplification. A target for 100 Gbps is BER below 10^-12 with sufficient margin to the forward error correction threshold.
• Run a dispersion check by measuring pulse broadening over the link length. Adjust dispersion compensation modules until the measured pulse width meets the target for the modulation format.

These checks require physical access to the fiber plant and permission to inject test traffic. Do not enable live traffic until BER and OSNR are within design targets.

Trade-off and limitation

WDM systems require precise wavelength alignment. Drift from temperature or aging lasers can cause channel overlap and increased error rates. Amplifier gain flatness must be maintained, otherwise some wavelengths suffer excess attenuation or noise figure degradation.

The cost of AWG or tunable laser modules is significant. Budget constraints often dictate DWDM versus CWDM more than pure technical merit. Also, WDM does not fix fiber loss or aging connectors. A clean plant is a prerequisite.

Actionable close

Map your distance, channel count, and growth horizon first. For 80 km with three 100 Gbps services and no growth, CWDM with passive mux and verified loss budget is a defensible low-cost choice. For growth beyond eight channels or need for amplification, start with DWDM and accept the higher capex for future-proofing. Verify with OSA, BER, and dispersion checks before production.