Opportunity
Submarine fiber optic cables carry over 99% of global internet data, yet designing these networks is extremely complex and costly. Cable costs increase with the number of optical fibers, and deployment expenses are influenced by seabed topography, ocean environment, and communication demands. Traditional planning approaches treat cable path planning and bandwidth allocation separately, often resulting in suboptimal networks that either overspend on high-capacity cables where not needed or fail to meet bandwidth requirements. Adding branching units (underwater splitters) can reduce cable length but introduces additional costs. Furthermore, real-world factors like earthquake risk zones, fishing activity, and varying seabed depths affect both laying costs and long-term maintenance expenses. There is a need for an integrated optimization method that simultaneously considers cable routing, branching node placement, and bandwidth capacity selection while accounting for geographic and operational risks.
Technology
This patent presents a system that models the submarine cable network optimization problem as a Weighted Edges Steiner Minimum Tree (WE-SMT) problem on the Earth's surface manifold. The processing module receives terminal node positions (e.g., cable landing stations) and bandwidth requirements between node pairs. It then determines possible distances and edge weights based on available cable specifications (each with a specific bandwidth capacity, e.g., 50 Tbps, 100 Tbps, and corresponding cost per kilometer).
The optimization module employs a novel WE-SMT algorithm that combines the Fast Marching Method (FMM) for geodesic path planning on triangulated terrain meshes with dynamic programming to simultaneously optimize: (1) geographical positions of Steiner nodes (branching units), (2) bandwidth capacity selection for each cable edge from discrete options, and (3) cable paths that minimize a total cost function including laying costs, Branching Unit (BU) costs (≈$1-3M each), and long-term repair risks (earthquake and human activity). The algorithm builds a directed acyclic graph (DAG) representing possible Steiner node positions, then traces back the minimum-cost tree. Higher-bandwidth edges are automatically shortened in the optimal solution, as they incur higher per-kilometer costs.
Advantages
- Integrated Optimization: Simultaneously optimizes node placement, bandwidth allocation, and cable routing, unlike traditional methods that handle them separately.
- Cost vs. Bandwidth Trade-off: Selects optimal cable specification (e.g., 50 Tbps vs 100 Tbps) per edge based on actual bandwidth demand, avoiding overspending.
- Geodesic Path Planning: Uses Fast Marching Method on real terrain data, accounting for seabed depth, earthquake zones, and human activity risks.
- Steiner Node Optimization: Dynamically determines optimal number and placement of branching units to minimize total network cost.
- Real-World Risk Integration: Includes earthquake-related and human activity-related repair rates in cost calculations.
- Demonstrated Savings: Testing showed the optimized topology reduced total network cost by 18.4% compared to uniform-bandwidth design (unoptimized: $162.50M; optimized: $132.50M).
Applications
- Submarine Cable Network Planning: Designing new transoceanic cable systems connecting multiple landing stations (e.g., Trans-Atlantic, Trans-Pacific).
- Telecommunications Infrastructure: Optimizing fiber-optic backbone networks crossing difficult terrain (mountains, deserts, or urban areas with varying right-of-way costs).
- Power Grid Cabling: Planning underwater or underground high-voltage power transmission lines with different capacity cables.
- Offshore Energy Networks: Designing cable systems for offshore wind farms, connecting multiple turbines to onshore substations.
- Disaster-Resilient Network Design: Routing cables around earthquake zones or high-fishing activity areas to minimize expected repair costs.
