Multi-landing submarine network architectures represent the pinnacle of optical networking engineering, combining sophisticated wavelength management, complex power distribution schemes, and advanced fault recovery mechanisms to deliver resilient, high-capacity connectivity across multiple geographic locations. Unlike traditional point-to-point submarine systems, multi-landing architectures leverage branching units (BUs) and wavelength management units (WMUs) to create intricate network topologies that optimize both capital and operational expenditures while providing unprecedented flexibility in traffic routing and network evolution.

Modern submarine systems have evolved from simple dual-terminal configurations to complex mesh networks connecting dozens of landing points through intelligent optical switching, reconfigurable add-drop multiplexing (ROADM), and adaptive power feeding architectures. These systems carry over 95% of global intercontinental data traffic, supporting multi-terabit-per-second transmission rates using dense wavelength-division multiplexing (DWDM) with coherent detection and advanced digital signal processing. The architectural decisions made during system design profoundly impact long-term network economics, operational flexibility, and service availability over the typical 25-year system lifetime.

This advanced technical analysis examines the theoretical foundations, implementation challenges, and optimization strategies for multi-landing submarine network architectures, targeting senior engineers and system architects involved in planning, deploying, or operating these critical global infrastructure systems.

1.1 Network Topology Theory for Multi-Landing Systems

Multi-landing submarine network topologies must balance competing objectives including path diversity, spectral efficiency, power budget optimization, and fault tolerance. The fundamental topological configurations include trunk-and-branch, ring, festoon, and hybrid mesh architectures, each presenting distinct trade-offs in terms of capital expenditure, operational complexity, and resilience characteristics.

Trunk-and-branch topologies dominate modern multi-landing deployments due to their optimal balance between fiber utilization efficiency and routing flexibility. In these configurations, a high-capacity trunk cable connects primary terminal stations while branching units provide optical and electrical distribution to intermediate landing points. The branching unit deployment strategy critically impacts system economics, with considerations including:

• Optical fiber pair allocation across trunk and branch segments based on traffic demand matrices
• Wavelength reuse optimization through reconfigurable optical add-drop multiplexing (ROADM)
• Power feeding topology selection (star, fishbone, branch-on-branch) to minimize high-voltage requirements
• Spectral guard band minimization using wavelength-selective switch (WSS) technology
• Network evolution planning to accommodate traffic growth and technology upgrades over 25-year lifetimes

1.2 Wavelength Management and Spectral Efficiency

Advanced wavelength management in multi-landing systems leverages reconfigurable optical filtering to maximize spectral utilization while maintaining operational flexibility. The wavelength selective switch (WSS) represents the enabling technology, providing software-controlled routing of individual wavelength channels or channel groups with sub-50 GHz granularity in modern implementations. This enables dynamic bandwidth allocation between trunk and branch segments without physical infrastructure modifications.

The spectral efficiency analysis for multi-landing networks must account for filter guard bands, which represent unused optical spectrum required to prevent inter-channel interference at optical add-drop boundaries. Fixed optical filter implementations typically require 100-200 GHz guard bands per filter transition, whereas WSS-based ROADMs can reduce this to 25-50 GHz through sharper filter roll-off characteristics. For a system with N branching points and M wavelength reuse instances, the total guard band overhead becomes:

Guard_Band_Total = N × M × GB_per_transition

Where:
  N = Number of branching units with OADM/ROADM functionality
  M = Average number of wavelength reuse transitions per branch
  GB_per_transition = Guard band width per filter transition (GHz)

Example:
  For system with 5 ROADM BUs, 3 reuse transitions each, 50 GHz guard bands:
  Guard_Band_Total = 5 × 3 × 50 GHz = 750 GHz

Note: WSS-based ROADMs reduce guard band requirements by ~50-75%
  compared to fixed filter implementations

1.3 Power Feeding Architecture and Voltage Budget Analysis

Power feeding represents a critical constraint in multi-landing system design, particularly for trunk-and-branch configurations with multiple power segments. The system voltage budget must accommodate the total power dissipation across all submerged repeaters, branching units, and cable conductor resistance while remaining within safety limits (typically ±15 kV shore-to-shore).

In multi-landing configurations with switchable branching units, the power feeding topology can be dynamically reconfigured to maintain service on unaffected segments during cable faults. Power feed branching units (PFBUs) contain high-voltage relays that can redirect feeding current paths, enabling scenarios such as:

• Single-end feeding of trunk segments while isolating faulted branch cables
• Dual-end feeding reconfiguration to restore powered segments after main trunk faults
• Branch cable repowering from trunk-side PFE after branch terminal failures
• Segmented power domain isolation for maintenance operations

V_total = Σ(N_rep × V_rep) + Σ(R_cable × I_feed × L_segment) + V_BU

Where:
  N_rep = Number of repeaters in powered segment
  V_rep = Voltage drop per repeater (~80-120V depending on capacity)
  R_cable = Cable conductor resistance (Ω/km)
  I_feed = Feeding current (typically 0.8-1.5A)
  L_segment = Length of cable segment (km)
  V_BU = Voltage drop across branching unit (negligible for modern designs)

Constraint: |V_total| ≤ V_max_safe (typically ±15 kV shore-to-shore)

Example calculation for 5000 km trunk with 3 branches:
  Repeaters: 100 × 100V = 10,000V
  Cable IR drop: 0.3 Ω/km × 1.0A × 5000 km = 1,500V
  Total: 11,500V < 15,000V ✓

2.1 ROADM-Based Branching Unit Architecture

Modern reconfigurable optical add-drop multiplexer (ROADM) branching units represent the most sophisticated wet plant equipment in submarine networks, combining optical switching matrices, wavelength-selective filters, and intelligent power management within titanium pressure housings rated for 8000-meter depths. The ROADM-BU architecture typically separates functionality between the branching unit proper (BU), which handles fiber pair routing and electrical power distribution, and the wavelength management unit (WMU), which implements optical filtering and channel manipulation.

The wavelength selective switch (WSS) forms the optical core of ROADM WMUs, enabling dynamic reconfiguration of optical spectrum allocation between trunk and branch paths. WSS devices utilize liquid crystal on silicon (LCoS), micro-electromechanical systems (MEMS), or planar lightwave circuit (PLC) technologies to route individual wavelengths or channel groups with software-controlled precision. Advanced WSS implementations provide:

• Channel granularity down to 6.25 GHz (flexible grid compatibility per ITU-T G.694.1)
• Per-channel optical power equalization through variable attenuation profiles
• Hitless wavelength switching with <50ms transition times
• Spectral shaping to compensate for amplifier gain tilt and filter concatenation effects
• Colorless operation allowing arbitrary wavelength assignment to any output port

The optical architecture within ROADM WMUs employs passive optical couplers to split incoming trunk and branch add signals, which then traverse wavelength-selective filter blocks before recombination. Each filter block can be implemented as:

• Fixed filters: Lowest cost and highest reliability, but inflexible spectrum allocation requiring pre-planning based on 25-year traffic forecasts
• Switched filters: Optical switches select between predetermined filter configurations (typically 4-6 states), balancing flexibility with reliability
• WSS-based filters: Fully reconfigurable with arbitrary spectrum allocation, highest flexibility but requires careful reliability engineering including WSS device protection switching

2.2 Trunk-and-Branch Network Topology Optimization

Optimizing trunk-and-branch network topologies requires simultaneous consideration of multiple competing objectives including minimization of capital expenditure (CAPEX), operational expenditure (OPEX), system latency, and maximization of network resilience and capacity utilization. The optimization problem becomes particularly complex in networks with numerous landing points and evolving traffic demands over multi-decade system lifetimes.

The trunk-and-branch topology selection process evaluates three primary variants:

• Star topology: Single trunk cable with branching units connecting directly to each branch station, minimizing trunk distance and latency but requiring larger fiber pair counts to accommodate all traffic flows
• Fishbone topology: Linear trunk cable with branch connections at intervals, optimizing fiber utilization through wavelength reuse but increasing path lengths for end-to-end traffic
• Branch-on-branch topology: Cascaded branching where branches further split into sub-branches, maximizing connectivity options in complex multi-tier networks but significantly increasing power feeding complexity and fault isolation challenges

2.3 Coherent Detection and Advanced Modulation for Multi-Landing Systems

Multi-landing submarine systems universally employ coherent optical detection combined with digital signal processing (DSP) to maximize spectral efficiency and transmission reach. Coherent receivers reconstruct both amplitude and phase information from received optical signals, enabling sophisticated modulation formats including dual-polarization quadrature phase shift keying (DP-QPSK), 16-ary quadrature amplitude modulation (16-QAM), and probabilistically-shaped 64-QAM (PS-64QAM) for metro-distance segments.

The modulation format selection in multi-landing architectures presents a challenging optimization problem because different system segments may experience vastly different optical signal-to-noise ratio (OSNR) requirements. Trunk segments connecting major landing points typically span thousands of kilometers through dozens of optical repeaters, demanding conservative modulation formats with high OSNR margins. Branch segments, particularly those connecting nearby islands or serving metro-reach applications, may traverse only a few hundred kilometers with minimal repeater counts, enabling higher-order modulations that deliver 2-3× improved spectral efficiency.

Modern submarine line terminal equipment (SLTE) designs incorporate "recoloring" nodes that enable modulation format conversion at cable landing stations. These nodes implement wavelength-switched mesh functionality, allowing signals arriving on trunk fiber pairs to be optically converted to different wavelengths and modulation formats for terrestrial distribution or branch cable propagation. The recoloring architecture provides:

• Modulation format adaptation matching transmission distance requirements (e.g., DP-QPSK for transoceanic trunk, DP-16QAM for regional branches)
• Wavelength grid conversion between fixed 50 GHz DWDM and flexible grid allocations
• Grooming of multiple lower-rate services onto single high-capacity wavelengths
• Protection switching and path diversity through wavelength-based routing

3.1 Wet Plant Reliability Engineering

Submarine equipment reliability requirements far exceed those of terrestrial optical networks due to the extreme difficulty and cost of undersea repairs. Modern repeatered submarine systems target mean time between failures (MTBF) exceeding 25 years for the wet plant as a whole, necessitating component-level failure rates below 100 FIT (failures in 10⁹ hours) for critical active elements like pump lasers and optical switches.

Branching units with reconfigurable optical functionality present particular reliability challenges because they incorporate active components (WSS devices, optical switches, photodiode monitors) within the harsh subsea environment. The reliability engineering approach for ROADM-BUs includes:

• Component derating with pump lasers operating at 50-60% of maximum rated power to extend service life
• Redundant optical switching matrices with automated failover to backup WSS devices upon performance degradation detection
• Hermetically-sealed optical subassemblies protecting sensitive components from hydrogen ingress and pressure cycling
• Extensive reliability testing including high-temperature operating life (HTOL), thermal cycling, and accelerated stress testing over 6-12 month validation periods
• Comprehensive design for manufacturing and assembly (DFMA) ensuring reproducibility across production runs

3.2 Fault Recovery Automation and Self-Healing Networks

Multi-landing networks with ROADM branching units can implement sophisticated automated fault recovery procedures that restore service on unaffected segments within seconds to minutes of fault detection. The fault recovery architecture operates at multiple layers:

Optical Layer Recovery: When optical power monitoring within ROADM WMUs detects a sudden loss of signal on trunk or branch input ports, the WMU can autonomously execute pre-programmed optical switching states to reload output paths from surviving inputs. For example, if the trunk input fails, the WMU switches to pass the full branch add spectrum to maintain loading on trunk output repeaters, preventing cascade failures from loss of optical power.

Electrical Layer Recovery: Power feed branching units (PFBUs) incorporate high-voltage relays that reconfigure power distribution topology upon detection of cable faults through power configuration analysis (PCA). The PCA algorithm continuously monitors feeding voltage and current to detect anomalies indicating shunt faults or cable cuts, then automatically switches relays to isolate faulted segments while maintaining power to operating equipment.

Network Management Layer Recovery: The submarine network management system (NMS) correlates fault indications from multiple sources including SLTE alarms, ROADM telemetry, and line monitoring equipment (LME) to determine fault location and execute coordinated recovery actions across multiple branching units. Advanced NMS implementations employ machine learning algorithms trained on historical fault data to predict optimal recovery configurations and minimize service disruption duration.

3.3 Line Monitoring and Fault Localization

Accurate fault localization in multi-landing submarine networks requires sophisticated line monitoring techniques that can pinpoint failure locations to within a few kilometers across systems spanning thousands of kilometers with multiple branching points. The primary monitoring technologies employed include:

Optical Time Domain Reflectometry (OTDR): Traditional OTDR systems inject narrow optical pulses into the fiber and analyze Rayleigh backscatter to create distance-resolved attenuation profiles. In multi-landing networks, optical filters within WMUs create monitoring segments that must be independently analyzed. Advanced correlation OTDR (COTDR) implementations use pseudo-random bit sequence (PRBS) coded pulses that provide improved sensitivity and noise immunity compared to simple pulsed OTDR.

High-Loss Loopback (HLLB) Monitoring: Each optical repeater incorporates a small-signal loopback path that allows in-service monitoring of repeater functionality without disrupting traffic channels. The HLLB signal traverses multiple repeaters before reflection back to the terminal, providing continuous health verification of the repeater chain. Changes in HLLB response characteristics can indicate degrading pump lasers, increasing span loss, or other pre-failure conditions enabling proactive maintenance.

Power Configuration Analysis (PCA): The PCA algorithm analyzes power feeding equipment voltage and current measurements to locate electrical faults including cable conductor breaks and shunt faults (short circuits to sea ground). By measuring voltage drop changes across different network segments and comparing to the predicted power budget model, PCA can typically localize faults to within 10-20 kilometers, guiding cable repair vessel deployment.

ΔL_min = (c × Δt) / (2 × n_fiber)

Where:
  ΔL_min = Minimum resolvable distance (meters)
  c = Speed of light in vacuum (3×10⁸ m/s)
  Δt = Optical pulse width (seconds)
  n_fiber = Refractive index of fiber core (~1.468 for SMF-28)

Example:
  10 nanosecond pulse width:
  ΔL_min = (3×10⁸ m/s × 10×10⁻⁹ s) / (2 × 1.468)
         = 3 / 2.936 = 1.02 meters

Trade-off: Shorter pulses improve resolution but reduce signal-to-noise ratio
  due to decreased pulse energy. Practical systems use 0.1-10 μs pulses
  providing 10-1000 meter resolution depending on dynamic range requirements.

4.1 OSNR Budget Engineering for Multi-Landing Systems

Optical signal-to-noise ratio (OSNR) budget analysis in multi-landing networks must account for amplified spontaneous emission (ASE) noise accumulation through the repeater chain, as well as additional penalties from optical filtering concatenation, chromatic dispersion, polarization mode dispersion, and fiber nonlinearities. The end-to-end OSNR budget establishes the fundamental limit on achievable transmission distance and spectral efficiency.

In trunk-and-branch configurations, different network segments experience distinct OSNR characteristics. Trunk segments accumulate ASE noise proportionally to the number of traversed repeaters (typically N_rep = 100-150 for transoceanic links), while branch segments may involve only a handful of repeater spans. This OSNR differential enables the use of reach-optimized modulation formats on different segments, with higher-order modulations (16-QAM, 64-QAM) reserved for short-reach branches where OSNR headroom exists.

The OSNR accumulation through a repeater chain with N amplifiers can be modeled as:

OSNR_total = P_sig / (N_rep × ASE_per_amp)

In dB units:
OSNR_dB = P_sig_dBm - 10×log₁₀(N_rep) - ASE_dBm_per_amp - 10×log₁₀(B_ref)

Where:
  P_sig = Signal power per channel (mW or dBm)
  N_rep = Number of amplified repeaters in cascade
  ASE_per_amp = ASE noise power added per amplifier
  B_ref = Reference bandwidth (typically 12.5 GHz for 0.1 nm)

ASE power per amplifier:
ASE_per_amp = (NF - 1) × h × f × B_opt × G

Where:
  NF = Noise figure of optical amplifier (typically 4.5-6.0 dB for EDFAs)
  h = Planck's constant (6.626×10⁻³⁴ J·s)
  f = Optical frequency (~193.1 THz for C-band)
  B_opt = Optical bandwidth (Hz)
  G = Amplifier gain (linear units)

Example for transoceanic system:
  N_rep = 150, P_sig = 0 dBm, NF = 5.5 dB, Span = 60 km
  OSNR_dB ≈ 0 dBm - 10×log₁₀(150) - 5.5 dB - (-58 dBm) = 30.7 dB

Critical threshold: Modern coherent systems require OSNR >15-20 dB for
  error-free operation depending on modulation format and FEC overhead

4.2 Capacity Planning and Traffic Matrix Optimization

Capacity planning for multi-landing submarine systems involves predicting traffic demand evolution over 20-25 year system lifetimes and optimizing fiber pair allocation, wavelength provisioning, and branching unit configuration to accommodate growth while minimizing capital expenditure. The traffic matrix defines expected bandwidth requirements between all landing point pairs, accounting for both direct point-to-point demands and transit traffic passing through intermediate nodes.

In ROADM-enabled networks, the wavelength allocation problem becomes a multi-dimensional optimization seeking to:

• Maximize wavelength reuse through intelligent spectrum partitioning between trunk and branch segments
• Minimize spectral guard band overhead by reducing the number of filter transitions
• Balance traffic loading across available fiber pairs to avoid creating capacity bottlenecks
• Provision sufficient dark fiber and spectrum reserves to accommodate future demand growth and technology upgrades
• Ensure resilience by maintaining adequate capacity margins for protection switching during fault conditions

The capacity planning process typically employs traffic forecasting models based on historical growth rates (compound annual growth rate or CAGR), market analysis, and consideration of new application drivers such as 8K video streaming, cloud gaming, and artificial intelligence workloads. Conservative submarine system designs assume 25-40% CAGR over the planning horizon, with lit capacity at system launch representing only 20-40% of ultimate equipped capacity to preserve upgrade headroom.

4.3 Nonlinear Impairment Management

Fiber nonlinearities represent a fundamental limitation on achievable transmission performance in high-capacity submarine systems, particularly for multi-landing architectures where optical signals traverse hundreds to thousands of kilometers through optically-amplified spans. The primary nonlinear effects include:

• Self-Phase Modulation (SPM): Intensity-dependent refractive index changes cause spectral broadening and pulse distortion within individual channels
• Cross-Phase Modulation (XPM): Intensity variations in one wavelength channel modulate the phase of co-propagating channels, causing inter-channel crosstalk
• Four-Wave Mixing (FWM): Nonlinear mixing between wavelength channels generates spurious frequency components that create noise and degrade OSNR
• Stimulated Raman Scattering (SRS): Power transfer from shorter to longer wavelengths causes tilt in the optical spectrum and channel-dependent power variations

Nonlinear impairment mitigation in multi-landing systems employs multiple strategies operating at system, line, and terminal equipment layers. System-level approaches include optimization of fiber type selection (large effective area fibers reduce nonlinearity), repeater span power budget design (maintaining optimal launch powers), and wavelength plan engineering (channel spacing and frequency allocation to minimize FWM products). Modern coherent transmission systems incorporate extensive digital signal processing that can partially compensate nonlinear distortion through techniques including digital back-propagation (DBP) and perturbation-based nonlinearity compensation.

5.1 Software-Defined Networking for Submarine Systems

Software-defined networking (SDN) principles are increasingly being applied to submarine cable systems, enabling programmable network control, automated service provisioning, and dynamic resource optimization across multi-landing architectures. SDN implementations in submarine networks leverage ROADM reconfigurability and intelligent terminal equipment to create flexible optical layer connectivity without manual wavelength planning or truck rolls for circuit provisioning.

Key SDN capabilities for multi-landing submarine systems include:

• Centralized network orchestration with topology discovery, path computation, and automated wavelength assignment
• Intent-based networking allowing service requests (e.g., "provide 400G between landing points A and C with <15ms latency") to be automatically translated into optical layer configurations
• Network telemetry and analytics providing real-time visibility into optical performance metrics (OSNR, BER, chromatic dispersion) across all network segments
• Closed-loop optimization algorithms that continuously adjust optical parameters (wavelength routing, power levels, dispersion compensation) to maximize network capacity and availability

5.2 Hollow-Core Fiber and Ultra-Low Latency Applications

Hollow-core fiber (HCF) technology, which guides light through an air-filled core rather than solid glass, offers the potential for ultra-low latency submarine transmission approaching the theoretical limit imposed by the speed of light in vacuum. Standard single-mode fiber exhibits a group refractive index of approximately 1.468, resulting in signal propagation at roughly 68% the speed of light in vacuum (c/n ≈ 204,000 km/s). Hollow-core fibers can reduce this to near n=1.0, achieving propagation speeds >99% of c.

For multi-landing submarine systems serving latency-sensitive applications such as high-frequency trading, cloud gaming, or time-critical scientific computing, the 30-35% latency reduction offered by HCF could justify deployment despite higher costs and technical challenges. A 10,000 km transoceanic HCF system could deliver signals approximately 16 milliseconds faster than conventional fiber – a significant advantage for applications where microsecond-scale latencies impact competitive positioning or user experience quality.

5.3 Quantum-Safe Cryptography Integration

The emergence of quantum computing poses a long-term security threat to conventional public-key cryptography algorithms (RSA, elliptic curve) used to encrypt submarine cable traffic. Post-quantum cryptographic (PQC) algorithms resistant to attacks by quantum computers are being standardized by organizations including NIST, with submarine cable operators beginning to evaluate integration strategies.

Multi-landing submarine architectures present unique challenges for quantum-safe cryptography deployment because branching units with ROADM functionality require access to wavelength channel information, potentially creating security vulnerabilities. Future ROADM-BU designs may incorporate quantum key distribution (QKD) capabilities or PQC-secured control plane mechanisms ensuring end-to-end encryption while maintaining optical layer flexibility.

5.4 Integrated Sensing and Communication

Submarine fiber-optic cables are increasingly being leveraged for environmental monitoring and early warning systems beyond their primary telecommunication function. Distributed acoustic sensing (DAS) and distributed temperature sensing (DTS) technologies use the fiber itself as a sensor, detecting vibrations and temperature changes along the entire cable length through analysis of optical backscatter patterns.

Applications include earthquake and tsunami detection, submarine landslide monitoring, marine mammal tracking, and climate research. Multi-landing submarine architectures with ROADM branching units can integrate sensing capabilities with minimal impact on communication capacity by dedicating specific dark fibers or wavelength channels to sensing applications, creating dual-use infrastructure that enhances both communication and scientific observation capabilities.

Multi-landing submarine network architectures represent the culmination of decades of optical networking innovation, combining sophisticated wet plant equipment, intelligent optical switching, advanced power management, and software-defined control to deliver resilient, high-capacity connectivity across global distances. The architectural decisions made during system design profoundly impact economics, flexibility, and operational characteristics over multi-decade lifetimes, requiring senior engineers to balance competing objectives including capital efficiency, spectral utilization, fault resilience, and technology evolution planning.

ROADM-enabled branching units have transformed submarine networks from static point-to-point links into dynamic mesh architectures capable of adapting to evolving traffic patterns through remote wavelength reconfiguration. Coupled with coherent optical transmission, advanced digital signal processing, and automated fault recovery mechanisms, modern multi-landing systems achieve service availability exceeding 99.99% despite traversing some of the harshest environments on Earth.

Looking forward, submarine network architects must prepare for capacity demands continuing to grow at 25-40% CAGR while incorporating emerging technologies including space-division multiplexing, hollow-core fiber, quantum-safe cryptography, and integrated sensing capabilities. The transition to software-defined networking principles promises to further enhance operational agility and enable new service models, ensuring submarine cables remain the indispensable foundation of global digital infrastructure for decades to come.

1. ITU-T Recommendation G.977.1, "Optically amplified WDM submarine cable systems," International Telecommunication Union, October 2020.
2. ITU-T Recommendation G.972, "Definition of terms relevant to optical fibre submarine cable systems," International Telecommunication Union, August 2024.
3. S. Mas Machuca et al., "Software-Defined Networking and Network Programmability for Next-Generation Submarine Networks," Journal of Optical Communications and Networking, vol. 13, no. 4, 2021.
4. Bergano, N.S. et al., "Submerged Plant Equipment," in Undersea Fiber Communication Systems, 2nd Edition, Academic Press, 2016.
5. Courtois, O. and Bardelay-Guyot, C., "Architectures and Management of Submarine Networks," in Undersea Fiber Communication Systems, 2nd Edition, Academic Press, 2016.
6. "Submarine Network ROADM Branching Units," Alcatel Submarine Networks Technical Documentation, 2020.
7. Antona, J.C., "The Submarine Network: A Complex Integrated System," SubOptic 2020 Conference Proceedings.
8. IEEE 802.3, "Ethernet Standard," Institute of Electrical and Electronics Engineers, ongoing development.
9. Rapp, L. and Eiselt, M.H., "Optical Amplifiers for WDM Submarine Systems," in Undersea Fiber Communication Systems, Academic Press, 2016.
10. "Global Submarine Cable Market Analysis 2024-2031," Market Research Reports, October 2024.