C-Band Classical Optical Communication Systems
- C-band classical communications line systems are frameworks transmitting multi-wavelength optical signals over fiber using advanced modulation, WDM, and spatial multiplexing.
- They integrate precise optical transmitters, multicore fibers, and coherent receivers to achieve aggregate capacities from sub-terabit to petabit scales with robust DSP and amplification schemes.
- Design trade-offs include managing dispersion, nonlinearity, and OSNR constraints while scaling performance through spatial and spectral multiplexing strategies.
A C-band classical communications line system is a physical and operational framework for transmitting multi-wavelength, high-capacity classical (non-quantum) optical signals over a fiber-optic medium confined to the C-band (conventionally 1530–1565 nm, but extended up to ~4.65 THz or more for advanced spatial multiplexing). These systems form the backbone of terrestrial, metropolitan, long-haul, and increasingly data center interconnect applications, engineered to reliably deliver aggregate capacities from sub-terabit up to petabit-per-second scale, typically using advanced modulation, amplification, wavelength-division multiplexing (WDM), and often spatial multiplexing.
1. System Architectures in C-Band Classical Line Systems
C-band line systems are organized around optical transmitters, fiber transmission media, and receivers—each with rigorously defined electronic and photonic sub-components.
High-capacity SDM-WDM coherent systems: The implementation in "389.3-Tb/s 1017-km C-band Transmission" (Kawai et al., 10 Dec 2025) employs 140 GBaud probabilistically shaped PS-36QAM on 4-channel CMOS DACs (32 GHz BW, augmented by 89 GHz doublers) paired with <10 kHz linewidth external-cavity lasers. A digital pre-emphasis—including DMUX and WSS-based flattener—equalizes transmitter frequency response.
Fiber: A 12-coupled-core G.652.D cladding multicore fiber with strong random inter-core coupling (<6 ps/√km SMD), α = 0.176 dB/km, β₂ ≃ –21.7 ps²/km at 1550 nm, and standard cladding geometry.
Receiver: A dual-polarization 70 GHz balanced photodiode plus 70 GSa/s oscilloscope feeds into offline (or real-time) DSP, performing CD compensation, 96×24 frequency-domain MIMO equalization, PDM demux, and soft-FEC-based PS-36QAM demapping.
IM/DD and DMT systems: In intra-datacenter or metro links, architectures such as the 216 GBd PS-PAM12 IMDD line (Nakamura et al., 23 Dec 2025) or 140 GBd OOK short-reach links (Ozolins et al., 2018) deploy high-speed InP-DHBT electronics, thin-film LiNbO₃ or EAM/MZM modulators, and high-BW PIN-PDs, often still relying on direct detection. DMT-based C-band line systems (Dochhan et al., 2020) interleave up to 8×50 GHz-spaced DWDM channels with VSB filtering via laser detuning for dispersion mitigation.
Hybrid and coexistence systems: Hollow-core fibers are engineered for low SRS/nonlinearity to enable co-propagation of C-band classical and QKD signals. These architectures use standard C-band ITU grid DWDM, modest EDFA gain stages, and sophisticated filtering (e.g., >95 dB notch at quantum channel) to enable error-free classical traffic alongside QKD (Honz et al., 2024, Honz et al., 2022).
2. Wavelength, Spatial, and Modulation Multiplexing Plans
C-band line system throughput is predominantly limited by available spectral and spatial resources:
WDM planning: Conventional systems use dense ITU grid spacing—e.g., 31 wavelengths × 150 GHz in (Kawai et al., 10 Dec 2025), or 17–25 channels × 100 GHz in HCF co-existence (Honz et al., 2022, Honz et al., 2024). Per-channel rates and required OSNR determine the optimal launch power and spacing for given reach.
Spatial multiplexing: Strongly coupled multicore fiber enables scaling beyond ~0.5 Pb/s aggregate by providing 12 independent cores, further enhanced by PDM to yield 24 spatial channels (Kawai et al., 10 Dec 2025). Achieving low spatial-mode dispersion and strong random coupling is critical to permit frequency-domain MIMO equalization.
Modulation formats: High-baud coherent systems now routinely use 140+ GBd PS-36QAM, achieving net rates >12 Tb/s/λ and >80 bit/s/Hz spectral efficiency (Kawai et al., 10 Dec 2025). IM/DD links explore 216 GBd (PS-PAM12), 140 GBd (OOK), and DMT for maximizing per-lane net rates and spectral utilization (Ozolins et al., 2018, Dochhan et al., 2020, Nakamura et al., 23 Dec 2025).
3. Link Budget, Optical Amplification, and OSNR Constraints
Accurate link budgeting is essential for system reliability, especially as spatial channels and WDM densities increase.
Per-span amplification and loss: In coherent SDM systems, each ~53.5 km span presents α·L ≃ 12.1 dB total loss; dual-stage EDFAs compensate with gain G ≃ 12 dB and NF ≃ 4.5 dB (Kawai et al., 10 Dec 2025). IM/DD and short-haul systems rely on single-stage EDFAs or preamps (NF ≈4 dB), often unamplified over <10 km (Ozolins et al., 2018, Nakamura et al., 23 Dec 2025).
OSNR targets: For high-order PS-36QAM at 140 GBd, minimum OSNR per 0.1 nm must exceed ~20 dB across all spatial and WDM channels (Kawai et al., 10 Dec 2025). IMDD and DMT formats typically demand OSNR in the 18–41 dB (0.1 nm) range, highly dependent on per-channel rate and reach (Nakamura et al., 23 Dec 2025, Dochhan et al., 2020).
Link budget formula:
with all terms in dB.
4. Dispersion, Nonlinearity, and Digital Processing
Chromatic and modal dispersion, as well as fiber nonlinearities, constrain achievable reach and aggregate rates.
Dispersion: For C-band, β₂ ≃ –21.7 ps²/km (SSMF), making high-baud and IM/DD links dispersion-limited at moderate distances. Techniques include pre/post optical filtering, RRC pulse shaping, and, in DMT, VSB filtering to eliminate power fading nulls via single sideband transmission (Nakamura et al., 23 Dec 2025, Dochhan et al., 2020). Dispersion-shifted fiber (DSF) with D ≈ 0 mitigates pulse broadening at extreme baudrates (Nakamura et al., 23 Dec 2025).
Nonlinear effects: In high-power, multi-core systems, SPM is limited by per-core launch power (~20 dBm), XPM is mitigated by random inter-core coupling and is averaged out by MIMO DSP, and SRS is essentially negligible in HCF by a factor of 35 dB relative to SSMF (Kawai et al., 10 Dec 2025, Honz et al., 2022, Honz et al., 2024).
DSP: Frequency-domain MIMO equalization scales as O(N²) for N spatial channels, e.g., 96×24 taps for full SDM-WDM-PDM operation, with sub-μs latency possible via sliding-windowed FFT (Kawai et al., 10 Dec 2025). DMT and IMDD links deploy FFE, DFE, and margin-adaptive algorithms for optimizing SNR per subcarrier or symbol (Nakamura et al., 23 Dec 2025, Dochhan et al., 2020).
5. Performance Metrics and Scaling Laws
Operational effectiveness is primarily measured by aggregate and per-wavelength net bitrates, spectral efficiency, reach, BER, and OSNR margin.
| Metric | Representative Value / Formula | Reference |
|---|---|---|
| Net bitrate per λ (multi-span SDM) | 12.55 Tb/s/wavelength (19×53.5 km) | (Kawai et al., 10 Dec 2025) |
| Aggregated net capacity | 389.3 Tb/s over 1017 km (31×12.55 Tb/s) | (Kawai et al., 10 Dec 2025) |
| Spectral efficiency (net, coherent SDM) | bit/s/Hz | (Kawai et al., 10 Dec 2025) |
| BER (IMDD, DMT, coexistence) | per channel (10 Gb/s OOK) | (Honz et al., 2024, Honz et al., 2022) |
| Maximum reach (coherent SDM) | 1017 km (19×53.5 km) | (Kawai et al., 10 Dec 2025) |
| Required OSNR (PS-36QAM) | dB/0.1 nm per channel | (Kawai et al., 10 Dec 2025) |
| IMDD net rate per lane | 582 Gb/s (216 GBd PS-PAM12, 11 km DSF) | (Nakamura et al., 23 Dec 2025) |
Performance scaling: By increasing spectral (C+L) and spatial (12→19 cores) resources, and baudrate (140→200 GBd), aggregate capacities extend well into the Pb/s regime (Kawai et al., 10 Dec 2025). In IMDD and OOK, reach is strictly limited by bandwidth and dispersion: e.g., 140 Gbaud OOK reaches ~5.5 km in uncompensated SSMF, extendable with DSF, integrated optics, and advanced DSP (Ozolins et al., 2018, Nakamura et al., 23 Dec 2025).
6. Practical Deployments, Co-Propagation, and Special Topics
Hollow-core fiber deployment: C-band classical systems are increasingly tested for quantum/classical coexistence. Low SRS and negligible XPM/FWM in HCF permit full-band DWDM (~17–25×10 Gb/s) with aggregate +11 to +12 dBm launch, co-propagating with a 1538 nm QKD channel. No measurable penalty is observed for the classical channels (BER < 10⁻⁹, open eyes) or key distillation at 330–660 bit/s (Honz et al., 2024, Honz et al., 2022).
Extended C-band in RF satellite systems: VSAT line systems operating in India's KPTCL network utilize 6.875–6.9465 GHz uplink and 4.650–4.7215 GHz downlink (extended C-band), with link budgets incorporating path loss, antenna gain (7.2 m hub, 1.2 m VSAT), and amplifier nonlinearities per the Saleh AM/AM–AM/PM model (Surekha et al., 2012). With careful phase/frequency compensation, these systems achieve BER ≪ 10⁻³ at modest 64–128 kb/s rates over geosynchronous links.
System connectivity and deployment: MCF-compatible SC connectors (<0.1 dB splice loss), WSS-based channel equalization, and field-installed cladding constraints drive practical architectural choices in SDM/WDM line systems (Kawai et al., 10 Dec 2025).
7. Design Insights, Trade-Offs, and Implications
DSP scalability: MIMO processing dominates coherent SDM architectures, with O(N²) complexity favoring frequency-domain implementations (Kawai et al., 10 Dec 2025). DSP in IMDD OOK/PAM links must balance equalizer depth with real-time constraints.
Shaping versus complexity: Probabilistic shaping (PS-36QAM, PS-PAM12) yields SNR gains (e.g., ~0.6 dB for PAM12) and higher net rates, trading off matcher memory and nonlinear predistortion effort (Kawai et al., 10 Dec 2025, Nakamura et al., 23 Dec 2025).
Reach-bandwidth trade-off: In DMT, bandwidth can flexibly scale channel count versus per-channel rate, enabling adaptive reach without changing core DSP (Dochhan et al., 2020).
Coexistence with quantum signals: Hollow-core fiber and strong filtering have enabled classical C-band WDM to coexist with DV-QKD on the same fiber, eliminating prior requirements for dark fiber and demonstrating no classical BER degradation at aggregate loads up to +12 dBm for 7.7 km (Honz et al., 2024).
Capacity scaling: Moving beyond 0.4 Pb/s field-tested capacity is feasible with C+L band extension, higher multicore counts, and 200+ GBd transceivers, but at the cost of more stringent dispersion, OSNR, and DSP requirements (Kawai et al., 10 Dec 2025).
A plausible implication is that continued progress in multicore fiber manufacturing, FEC, and DSP architecture will determine the longevity and ultimate scaling limits of C-band classical communications line systems.