Country-Pair Analysis

Creates a virtual sourceconnected to all of Country A's entry nodes and a virtual targetconnected to Country B's mainland terminal nodes. This allows multi-source BFS to find the best path from any entry point in A to any mainland terminal in B.

The source country may start from any valid entry node (all LPs + DBUs). The target country only accepts mainland terminals (mainland DBU + mainland LPs with landmassGroup === 0). Offshore islands in the target country (e.g., India's Andrott/Amini) are excluded. Once a path reaches any mainland target terminal, it stops rather than continuing across the target country's domestic network.

The mainland heuristic assumes the landmassGroup === 0 landmass is the primary destination. For countries where the relevant target is actually an offshore territory, this may exclude valid paths. If no mainland LPs exist, it falls back to all LPs.

Enumerates paths between the two countries, ordered by cable hop count(number of distinct cables on the path). Domestic edges do not count as hops. Multiple submarine edges from the same cable count as 1 hop.

Uses a bucket-based BFS where buckets are indexed by cable hop count. Within each hop level, all node-to-node expansions happen before moving to the next hop. This ensures paths are found in cable-hop order even when a single cable traverses many BU nodes.

• 500 paths maximum per query — may miss rare alternative routes in dense networks
• 50,000 iteration cap — guards against combinatorial explosion but may return partial results
• Paths are unweighted — a short cable hop and a trans-oceanic hop count equally

Weighted shortest paths (distance, latency), Yen's K-shortest-paths for diversity, constrained path enumeration (max hops, must-include/exclude cables).

Filters the enumerated paths to show route diversity. Cables used in shorter-hop paths are excluded from longer-hop paths. Only paths that introduce at least one new cable (not already seen at a lower hop level) are kept.

Paths are grouped by hop count and processed in ascending order. A cumulative set of "seen cables" is maintained. At each hop level, a path is included only if it contains at least one cable not in the seen set. After processing a hop level, all its cables are added to the seen set.

This heuristic measures cable diversity, not physical route diversity. Two cables that share the same physical corridor for most of their length are treated as fully independent alternatives. A 3-hop path that uses one completely new cable is preferred over a 2-hop path that reuses cables from the 1-hop level, even though the 2-hop path might be more resilient.

Physical corridor overlap detection, SRLG (Shared Risk Link Group) aware filtering, topological diversity scoring (number of distinct geographic regions traversed).

• None (baseline) — unrestricted pathfinding, all routes considered
• No-transit — blocks all LP and DBU nodes of third countries. Only direct cable connections between the two countries are considered. Useful for understanding direct connectivity without relying on intermediate countries.
• Custom— selectively block specific cables, landing points, countries, or individual cable segments. Enables "what-if" scenario analysis.

Blocked countries: All LP and DBU nodes for that country are removed. Blocked cables: All submarine edges for that cable are removed. Blocked LPs: The specific LP node is removed. Blocked segments: Individual submarine edges are matched by their endpoint coordinates within ~0.01° tolerance (only first/last points checked for efficiency).

• Line-segment blocking uses endpoint matching only with ~0.01° tolerance — may fail to match segments with different coordinate precision
• No-transit mode blocks at the node level — it cannot express "allow transit through Country X but not Y"
• Custom mode's robustness comparison re-runs baseline pathfinding on every mode change

Countries (other than A and B) whose removal would break all paths between A and B.

First, find baseline paths and extract transit countries (middle nodes in path sequences). For each transit country, simulate blocking all its nodes and re-check reachability. If A can no longer reach B, that country is "critical transit."

Only tests transit countries found in the first 5 baseline paths. A critical transit country that only appears on paths beyond the 5 shortest may be missed.

Each path's submarine edges carry their original geographic coordinates from cable geometry. Consecutive submarine edges from the same cable are merged into continuous segments. Between submarine segments (across domestic gaps), node coordinates are interpolated to fill the gap. Longitudes are normalized to a continuous range (cross-dateline support). Paths are then color-coded by hop level on the map.

Domestic gaps use straight-line interpolation between nodes, not actual terrestrial routes. This can produce visually misleading straight lines across land masses (e.g., a domestic gap across a continent).