Submarine Cables

Submarine Cable Fault Tracker

A live world map of submarine cable systems with auto-detected fault status. Dual-trigger detection (news + country-level traffic anomaly) keeps the state honest; AI extraction filters out false positives; repair progress and recovery are inferred from traffic baselines.

Click any highlighted cable to inspect the underlying evidence chain. Drag the timeline to replay the last 90 days.

Loading… Normal Faulted In repair Long-overdue
Now

Submarine Cable Fault Tracker

Live world map of submarine cable status — faulted, in repair, recovered. Dual-trigger detection from news and country-level traffic anomalies, with AI-assisted news extraction.

Category
submarineCables
Tags
submarine-cable, outage, telegeography, cloudflare-radar, ioda, maplibre, fault-tracking

Inputs

  • None — interactive map of all known submarine cables

Outputs

  • Color-coded cable status (green / red / yellow / stale)
  • Click-through evidence chain: news links, traffic anomaly z-scores, AI summary
  • Estimated repair ETA based on regional historical P50
  • Timeline scrubber to replay the last 90 days

Use Cases

  • First-look situational awareness when a regional outage is reported
  • Pre-NOTAM check before a planned maintenance window in a region with active faults
  • Explaining cable health to non-engineering stakeholders

Data Sources

  • TeleGeography submarinecablemap.com (cable geometry)
  • Cloudflare Radar (country traffic anomaly)
  • IODA / Georgia Tech (country signal cross-check)
  • Google News RSS (fault keyword search query)
  • SubTel Forum global RSS (https://subtelforum.com/feed/)
  • SubTel Forum Cable Faults & Maintenance category RSS (https://subtelforum.com/category/cable-faults-maintenance/feed/)

Limitations

  • TeleGeography geometries are illustrative, not real routes — fault marker is regional, not precise
  • Country-level traffic anomalies are not always cable-attributable (DDoS, blackouts, policy events)
  • Repair-vessel AIS is reserved for a later phase; in-repair state currently uses a 3-day fallback rule
  • Recovery detection requires ≥7 consecutive days of normal traffic baseline