Subsea Cables and Network Power

By the end of this lesson, you will be able to...

• Trace how traffic reaches another region through cable routes and landing points.

• Identify when apparent redundancy shares a physical or legal dependency.

• Explain how outage, repair, jurisdiction, and route choice change network power.

Idea in one sentence: Global connectivity depends on physical routes that can carry traffic, fail, be repaired, and fall under different jurisdictions.

Core Insight

Imagine this at 09:10 on Monday: a clinic worker on the island tries to open a patient's record hosted in a foreign cloud region. The request normally returns in under a second. Today it waits, then fails. The clinic's local network is working. The cloud region is working. The missing part is the international path between them.

The worker does not need a theory of global networking to notice the pressure. A record that exists but cannot be reached is unavailable to the person who needs it.

An island government launches a digital public-service portal. Its cloud region is abroad, but citizens can reach it through two listed internet providers. The dashboard says the service is redundant.

Then an undersea cable fault cuts one route. The second provider still sells connectivity, yet most of its international traffic uses the same landing station and the same remaining cable system. The portal works slowly, video consultations fail, and backup data transfers wait.

The internet is not a cloud drawn above a map. Long-distance traffic travels through fiber, cable landing stations, terrestrial backhaul, routers, power systems, and repair operations. A service is resilient only when its alternatives avoid the dependency that actually failed.

The Small Situation

The island's public portal has three important flows:

• citizens send requests to the cloud region;
• clinics upload records each night;
• the government replicates backups to another region.

The operator assumes that buying two provider contracts solves the connectivity problem. That is a reasonable first model because the contracts have different names.

It breaks when both providers converge at the same landing point. The contracts are commercially separate but physically correlated. A damaged cable, local power failure, construction accident, or access restriction at that point can affect both.

The missing model is a route map: not merely who invoices the customer, but which physical paths, facilities, operators, and jurisdictions carry the traffic.

The Moving Parts

Plain meaning:

A subsea cable is a physical fiber route across water. It connects to land at a landing station, then continues over terrestrial networks toward data centers and users.

In the island case:

A citizen's request travels from a local access network to a landing station, through a cable, to another landing station, then across terrestrial networks to the cloud region. The response returns through one or more paths.

Technical name:

This is a network path. A path has capacity, latency, ownership, failure modes, and a repair process. Several logical services can share one physical path without making that sharing obvious.

A Worked Trace: Two Providers, One Failure Domain

Start with the portal's normal request.

The naive conclusion is: "We bought two providers, so we have two independent paths."

The stronger conclusion is: "We have two commercial paths, but perhaps one physical failure domain." The difference matters because redundancy works only when the alternate route can carry the needed traffic after the relevant failure.

Now change one variable. Suppose the island adds a second cable landing station on another coast, with terrestrial backhaul that does not share the first station's power or road access. This costs money and may add latency for some users. It reduces correlation: one local incident is less likely to remove both international exits.

So far, the mechanism is visible. Traffic is not simply "online" or "offline." It follows routes with finite capacity and shared dependencies. A reroute can preserve basic availability while increasing latency, congestion, cost, or exposure.

Repair, Jurisdiction, and Leverage

Cable faults are not automatically geopolitical events. Fishing activity, anchors, equipment failure, weather, and accidents can all damage infrastructure. But physical geography creates leverage because repair and operation require access to specific places.

Repair often involves locating the fault, sending a specialized vessel, recovering cable, splicing it, testing it, and returning it to service. Permissions, weather, port access, security conditions, and the availability of repair vessels can change the recovery time. A route crossing several jurisdictions can therefore have different operational and political risks from a route with fewer or more cooperative access points.

Jurisdiction also matters before a failure. Landing stations, permits, local power, data rules, lawful interception regimes, ownership rules, and procurement decisions shape who can operate infrastructure and under what conditions. This does not mean that every jurisdiction can read every packet or control every route. It means the route map should include the institutions that can affect facilities, operators, and access.

Trade-offs and Limits

The central trade-off is between efficiency and resilience. Concentrating traffic on a small number of high-capacity routes can lower cost and simplify operation. It can also create shared failure domains. Diverse routes, landing stations, and backhaul paths reduce correlation, but cost more, may be slower, and require coordination across more operators.

What improves: A route map reveals whether a backup is real, identifies where congestion will appear after a fault, and gives operators a basis for prioritizing critical traffic.

What becomes more expensive: Independent paths require additional contracts, equipment, peering, capacity reservations, monitoring, and sometimes separate landing infrastructure.

What can still fail: Paths that look independent can share a cable corridor, power system, port, cloud region, or political risk. A second route may have insufficient spare capacity during a major failure.

What this does not solve: Cable redundancy does not replace application-level resilience, data backups, incident communication, or cloud-provider diversity.

Signals to watch: route diversity, landing-point concentration, utilization after reroutes, packet loss, latency, repair estimates, permit delays, and capacity reserved for critical workloads.

What an incident team should record

When international connectivity degrades, an operator should record the affected prefixes or services, the time routing changed, the observed latency and packet loss, the remaining capacity, and which physical facilities the alternative paths share. This avoids a familiar mistake: declaring a route independent because traffic still flows. The test is whether the path remains usable for the service promise after the failed dependency is removed.

For the island portal, clinics may receive priority while nightly backup replication pauses. That is an operational choice, not proof that the network has recovered. The system is deliberately trading freshness of backups for access to patient records. Making that choice explicit helps decision-makers see what the remaining route actually buys them.

Common Confusions

Confusion: Two providers always mean two independent routes

Why it is tempting:

Two contracts and two logos are visible.

Better model:

Trace physical cable systems, landing stations, terrestrial backhaul, and shared capacity. Commercial diversity can hide physical concentration.

Confusion: A cable outage means total disconnection

Why it is tempting:

Failures are often described as a cable being "cut."

Better model:

Traffic may reroute. The useful question is what capacity remains, which workloads are degraded, and how long repair takes.

Check Your Understanding

Check: Two routes leave a country through different providers but terminate at the same landing station. What should the analyst investigate next?

Think first, then reveal.

Answer: Investigate whether they share cable segments, power, terrestrial backhaul, access roads, repair dependencies, or jurisdictional exposure. The landing station may be a shared failure domain.

Check: A backup route adds 40 ms of latency but remains available during a cable fault. Is it useless?

Think first, then reveal.

Answer: No. It may protect critical services while lower-priority work is delayed. The decision depends on the workload's latency tolerance and the capacity of the backup route.

Practice

A country has three cable systems, but all land at one coastal district. A proposed fourth cable lands on a different coast and uses a separate terrestrial route, but costs more and has slightly higher latency. Review the proposal.

A good answer should identify the existing landing-point concentration, explain why the fourth cable reduces a shared failure domain, and name the trade-off in cost and latency. It should also ask whether the new route has independent power, repair access, sufficient capacity, and a different cloud or upstream dependency.

Resources

• [REFERENCE] Submarine Cable Map — Focus: Inspect how cable systems, landing points, and routes make network geography visible.
• [REPORT] The Internet in the Gulf — Focus: Compare physical cable geography with concentration and resilience questions.
• [REFERENCE] International Cable Protection Committee — Focus: Learn how cable protection, repair, and marine activity affect operational resilience.

Key Takeaways

• Long-distance connectivity depends on physical cable, landing, backhaul, and repair systems.
• Commercially separate providers can share one physical failure domain.
• Redundancy is useful only when the alternate path can carry needed traffic after the relevant failure.
• Repair time and jurisdiction are part of network power because they shape access to critical places.
• A route map should show capacity, shared dependencies, substitutes, and the decision affected by disruption.