Starlink and Cellular Hybrid Connectivity for Vessels

The Starlink terminal on a North Sea supply vessel stopped passing traffic at 06:40 on a Tuesday. There was no alarm and no obvious fault. The terminal was powered, the status light was normal, and the link was simply not there. The bridge noticed when the weather routing did not refresh. The engine room noticed when the overnight diagnostic file was still sitting in the queue. Shore noticed nothing at all, because a vessel that has gone quiet looks exactly like a vessel with nothing to report.

The link came back forty minutes later. Nobody could say why it went, and nobody could promise it would not happen again. That is the position most operators end up in after installing a satellite terminal as a single path: an excellent connection, right up to the moment it is not there, with no second answer on board.

This article is about the second answer. It leaves the cost comparison between the technologies to a separate article and concentrates on how to run satellite and cellular together on one vessel: which architecture fits which operating profile, how the switchover actually behaves, which traffic belongs on which path, and how to stop a standby link from turning into a standing cost.

Why one satellite path is a single point of failure

A low earth orbit terminal makes a strong primary connection. It is also one antenna, one power feed, one service account and one view of the sky, and each of those is a way for the link to stop. The failure modes are ordinary ones:

• Obstruction. Alongside a quay, under a gantry crane, in a lock, or in the shadow of a high superstructure, the terminal loses the clear sky it depends on. This happens most often in port, which is exactly where the vessel has the heaviest administrative traffic: clearance documents, loading bills, crew changes and shipping instructions.
• Weather. Heavy rain and dense cloud degrade the link, and a squall does not check whether the vessel is mid-upload.
• Hardware and power. A terminal, a cable run or a power supply that fails takes everything with it when there is nothing else on board.
• Congestion and prioritisation. On capped or lower-priority plans, throughput in a busy area can fall well below the datasheet figure.
• Service and licensing gaps. Coverage and permission to operate are separate questions, and both vary by water.

Taken together, these are arguments for adding a second path. Satellite remains a strong primary. What matters is that these failure modes are largely uncorrelated with terrestrial mobile coverage: a blocked sky view, a rain cell or a terminal fault does not affect the coastal cell towers the vessel can also reach. That lack of correlation is what gives a hybrid its value, and it is why cellular makes a stronger second path than a second satellite subscription.

The two hybrid architectures: active-passive and active-active

There are two ways to run two paths on a vessel, and the difference has practical consequences. It determines how much of the year each link carries, how failover is configured, and what the connectivity actually costs.

Weconnect describes both models on its Starlink page: a pay-per-GB 4G/5G SIM on standby behind a satellite primary, and an active-active configuration where cellular carries real traffic alongside it. The choice between them follows from where the vessel actually operates. A vessel that spends most of its operating hours within reach of coastal towers is throwing away the cheaper, lower-latency path if it leaves cellular idle. A vessel on deep-water routes is right to keep cellular as insurance.

Why the backup path should be multi-network

A backup connection tied to one mobile operator inherits that operator’s coverage map, and a vessel does not stay inside one coverage map. It crosses territorial waters, calls at ports in different countries, and works coastlines where one carrier is strong and the next is weak. A single-carrier SIM in the standby router performs well in its home market, and its usefulness declines as the vessel moves outside that operator’s footprint, which is frequently where the backup is needed most.

A Non-Steered Multi-Network-SIM removes that dependency. Non-steered means the device selects the strongest available network at its position instead of being pushed onto whichever partner network the provider prefers. Weconnect provides this across 700+ carrier partnerships in 195+ countries, so the standby link can draw on the best available signal at the vessel’s position. For a hybrid, that is what keeps the second path usable across the full operating area.

Pay-per-GB is what keeps a standby link alive in the budget

A standby connection has an awkward property: it costs money and, if everything goes well, it does nothing. Backup links billed at a fixed monthly rate per vessel are the first line questioned at a budget review, and on a fleet of fifteen vessels the annual figure is large enough that somebody will question it.

Pay-per-GB billing changes that conversation. The IoT-SIM-Implementierung generates cost when data flows and close to nothing when it does not, which makes an idle standby path cheap to keep and removes the incentive to cancel it. It also makes the link cheap to test, and testing is the part most operators skip.

A backup that has never been exercised on a specific vessel, with that vessel’s router, firmware, SIM and antenna, remains an assumption about how the equipment will behave under a fault it has never seen. A scheduled test, disconnecting the satellite path in controlled conditions and confirming that traffic moves and comes back, converts the assumption into a known quantity. Run it at commissioning and at least once a year after that.

What actually happens at the moment of switchover

Two configuration decisions determine whether failover works in practice.

Detection: how the router knows the link is gone

The simplest failover triggers on link state: the interface goes down, the router switches. That catches a terminal losing power or a cable coming loose. It does not catch the more common failure, which is the one in the opening of this article: an interface that is up and passing nothing. For that, the router needs active health checks, meaning it sends small probes to a known destination and treats sustained failure as a fault regardless of what the interface reports. Without health checks, a vessel can sit on a dead primary link for as long as it takes a person to notice.

Behaviour: what breaks and for how long

Switchover is quick, and it is still noticeable on board. The public IP address changes, so anything with an established session is affected. VPN tunnels re-establish, remote desktop sessions drop, VoIP calls end, and file transfers restart unless the protocol supports resuming. Systems that reconnect automatically recover in seconds. Systems that require somebody to press a button stay down until they get it. Listing per vessel which onboard systems fall into which category is worth doing before the first real failure answers the question for you.

Set the failback rule as deliberately as the failover rule. A primary link that recovers intermittently will make an eager router flap between paths, breaking sessions each time. Requiring the primary to be stable for a defined period before traffic returns to it costs a few minutes of the cheaper path and prevents a great deal of instability.

Traffic policy: decide what belongs on which path

The most common design error in a vessel hybrid is treating the second path purely as an emergency spare. Some traffic is better placed on cellular permanently, even when the satellite link is healthy, because it is small, critical, and expensive to lose.

That last row is the one that produces surprise invoices. When the primary fails at 21:00 and the whole vessel rolls onto a metered cellular link, crew streaming is what consumes the allowance, while the operational systems barely register. The mechanism for preventing it is the same segmentation used for separating crew and operational traffic on a single link: VLANs on the marine router, with quality-of-service rules and a crew cap that applies specifically when the vessel is running on the fallback path.

Nearshore, the second path does most of the work

For coastal and short-sea operators, calling cellular the backup understates what it does. With a high-gain broadband antenna and an amplifier, maritime Internetlösungen reach up to 70 km from the coast, against roughly 8 km on standard equipment, delivering up to 150 Mbit/s down and 50 Mbit/s up depending on the network. In port, alongside, in locks and on inland waterways, the cellular path is frequently the stronger of the two, and it is available in precisely the conditions that obstruct a satellite terminal.

For those vessels the honest description of the architecture is active-active, with cellular carrying real traffic every day. Which of the two should be primary for any given fleet is a cost and coverage question, worked through in detail in the cost comparison of 4G, Starlink and VSAT. The point here is narrower: a hybrid earns its cost when the routing policy reflects where the vessel really operates.

Running a hybrid across a fleet

One vessel with two links is a configuration. Fifteen vessels with two links each is an operational system, and it introduces a failure mode that does not exist on a single vessel: the silent failover. The hybrid works exactly as designed, the vessel drops onto cellular, traffic continues, nobody on board notices anything, and the primary link stays broken for eleven days until the data invoice arrives. Redundancy that hides faults is only half of the job.

A Konnektivitätsmanagement-Plattform closes that gap. Per-SIM usage visible in real time turns an unexplained jump in cellular consumption on one vessel into an alert on the same day, well before it appears as a line item at month end. Data caps and threshold alerts bound the exposure when a fallback runs longer than expected. Cost-centre allocation puts the consumption against the right vessel or project. SIMs can be provisioned before a vessel enters service and suspended when it leaves, and pooled allowances let a fleet absorb one vessel’s bad month without changing every plan.

What a working hybrid setup looks like

• Above deck: satellite terminal sited for the clearest practical sky view, and a high-gain maritime 4G antenna mounted as high as the structure allows, kept clear of the terminal and of radar.
• Below deck: a marine-grade router that supports multiple WAN interfaces, health-check based failover, VLANs and quality-of-service rules. The router is where the hybrid is actually built.
• Konnektivität: a non-steered multi-network SIM on pay-per-GB billing for the cellular path, alongside the satellite subscription.
• Policy: written down per vessel. Which path is primary, what triggers failover, how long the primary must be stable before failback, which traffic is pinned to cellular, and what the crew cap is on the fallback path.
• Monitoring: both paths visible from one platform, with alerts on unexpected cellular consumption so a silent failover is caught in a day.
• Testing: a controlled failover test at commissioning and annually, recorded per vessel with the recovery time and any system that needed manual intervention.

Häufig gestellte Fragen

Do I still need a 4G SIM if my vessel has Starlink?

Yes, if the vessel cannot afford to be offline. A satellite terminal is a single path that can be blocked by a quay crane, degraded by heavy weather, or taken out by a hardware or power fault. A multi-network cellular SIM provides a second path whose failure modes are unrelated to the first. On pay-per-GB billing it costs close to nothing while it sits on standby.

What is an active-active setup on a vessel?

Active-active means both the satellite and the cellular link are up at the same time, with traffic distributed across them by policy. It suits vessels that spend significant time within coastal cellular coverage, where the cellular path is often faster and cheaper. Active-passive keeps cellular purely on standby and suits offshore profiles.

How does the router know when to switch from Starlink to cellular?

It depends on how failover is configured. Link-state detection switches when the interface physically goes down, which catches power and cable faults. Health-check detection sends small probes to a known destination and switches when they fail, which also catches a terminal that is powered and connected but passing no traffic. Health checks are the more reliable trigger of the two.

Does a backup SIM cost money when it is not being used?

On a pay-per-GB model, close to nothing. The SIM generates cost when data flows and stays effectively dormant when the primary link is healthy, which is what makes an always-available standby path affordable to keep and cheap to test. Fixed monthly backup contracts have the opposite behaviour and tend to get cancelled.

Which traffic should stay on cellular even when Starlink is working?

Small, critical flows: position reporting, alarms, engine and sensor telemetry, and remote access for support. They consume very little data and are the traffic you least want to lose, so keeping them on a second path means a satellite fault does not blind the shore team. Bandwidth-heavy traffic such as crew welfare and large operational uploads belongs on the primary path, with a cap applied when it falls back.

Nächste Schritte

Weconnect supplies the cellular half of a hybrid vessel setup: non-steered multi-network SIMs across 700+ carrier partnerships in 195+ countries, pay-per-GB billing that keeps a standby path affordable, long-range maritime antenna and router hardware where it is needed, and a single platform showing every SIM in the fleet. Tell us where your vessels operate and what is on board today, and we will work through the architecture, the routing policy and the failover behaviour with you. Challenge us with your connectivity requirements. Direct response within 4 business hours.