Insights

How to design a superyacht network that never drops guest Wi-Fi

Guest Wi-Fi depends less on access points than on network design. How segmentation, multi-WAN routing and roaming make link failures invisible to the people on board.

8 November 2026  ·  7 min read

Guest Wi-Fi that never appears to drop comes from network design, not from expensive access points. Three things matter most: strict VLAN separation, multi-WAN failover that moves traffic between VSAT, Starlink, cellular and shore links without breaking sessions, and roaming engineered for a steel-and-glass hull. With the routing and segmentation right, modest hardware performs well. With them wrong, no amount of hardware will compensate.

No link at sea is up all the time. The realistic goal is that individual links fail, sometimes several times a day, and guests never notice. What follows sets out the design choices that make that possible and the questions worth asking at new-build or refit stage.

Why routing matters more than hardware

Most guest complaints blamed on ‘the Wi-Fi’ are not radio problems. Common examples include:

  • A video call that freezes when the yacht switches from Starlink to VSAT.
  • A streaming service that rejects a new public IP address.
  • A phone that stays connected to an access point two decks away.
  • Crew backups that saturate the uplink during dinner.

Each of these comes down to a routing, policy or roaming decision.

Enterprise-grade access points and switches are necessary, but they are the easy part to specify. A network that feels seamless differs from one that generates calls to the ETO in three areas: how traffic is segmented, how uplinks are chosen and monitored, and how clients move between access points. That is where design time should go.

VLAN separation: give every system its own lane

A superyacht network carries traffic with very different priorities and risk profiles over shared infrastructure. Separating it into VLANs, with firewall rules between them, is the foundation for both performance and security. A typical marine network segmentation includes:

  • Guest: internet only. Client isolation stops guest devices from seeing each other or anything else on board.
  • Crew: internet and limited internal services, with bandwidth policies separate from guests.
  • Bridge and operational technology: tightly restricted, with no route to guest or crew networks.
  • AV and control: entertainment, lighting, blinds and HVAC interfaces, reachable only from approved control devices.
  • Owner and family: often a distinct segment with higher priority.
  • Management: switch, access point and firewall administration, isolated from everything else.

Segmentation contains faults as well as threats. A misbehaving smart TV or an infected guest laptop stays inside its own segment rather than affecting the whole vessel. Segmentation also allows per-VLAN traffic policies, so the guest network can be guaranteed a share of the uplink whatever crew or onboard systems are doing. Flag, class and insurers increasingly expect cyber risk management, and a clearly segmented network is far easier to demonstrate and audit.

One practical caution applies. Casting and discovery protocols such as AirPlay and Chromecast do not cross VLAN boundaries by default. They need deliberate handling, such as multicast DNS gateways or tightly scoped rules, so guests can use cabin screens without undermining the segmentation.

Multi-WAN failover: plan for every link to fail

Most yachts now carry several uplinks, each with different strengths:

  • Starlink Maritime (LEO): high throughput and low latency. It is subject to obstruction from masts, radar arches and superstructure, and to brief interruptions.
  • VSAT (typically GEO): predictable coverage and, depending on the service plan, committed bandwidth, with higher latency.
  • 4G/5G cellular: excellent near the coast and in port, unavailable offshore.
  • Shore connection: marina fibre or Ethernet when alongside.

Multi-WAN design decides how these links work together. There are three main approaches:

  1. Failover: one primary link, with others on standby. It is simple, but sessions usually break on switchover because the public IP address changes.
  2. Load balancing: sessions are spread across links. Capacity is used better, but each session is still tied to one link and drops if that link fails.
  3. Bonding via a shore-side endpoint: traffic is tunnelled over all available links to a server ashore, which presents a single stable public IP. If a link fails mid-session, packets continue over the others, so calls and VPNs survive.

For guest Wi-Fi that genuinely feels continuous, bonding is usually the right answer for at least the guest and owner segments. Streaming services also see a consistent location. Bonding adds overhead, and the aggregation endpoint must itself be resilient, so treat it as a design decision rather than a default.

Failover details that matter

  • Health checks beyond link state. A modem can report ‘up’ while passing no traffic. The network should monitor real reachability, latency and packet loss, and act on those measurements.
  • Hysteresis. A link should be stable for a set period before traffic returns to it. This prevents flapping during intermittent obstruction.
  • Policy-based routing. Traffic can be steered by segment and application. Latency-sensitive guest traffic goes on the fastest path, bulk updates and backups on the cheapest or least loaded link, and critical operational traffic on the most predictable one.
  • Cost awareness. Metered links should be used deliberately, never as an accidental default after a failover.

Roaming: engineering for steel, glass and stairwells

Yachts are hostile radio environments. Metal bulkheads, coated glazing and dense joinery weaken signal unpredictably, and guests move constantly between cabins, decks, the beach club and the tender garage. Seamless yacht Wi-Fi roaming depends on several factors:

  • More access points at lower power. Coverage should be designed space by space, with moderate transmit power, so clients see one clear best access point rather than several marginal ones.
  • Survey and validation. Predictive designs from drawings are a starting point. Coverage must then be validated in the finished interior, after joinery and before delivery.
  • Centralised control. A controller or cloud-managed platform should coordinate channels, power and client steering across the vessel.
  • Fast roaming standards. 802.11k, v and r help devices move quickly between access points. Some devices handle 802.11r poorly, so it should be tested with the devices guests actually bring.
  • Band steering and minimum signal thresholds. Capable clients should be moved onto 5 GHz or 6 GHz, and devices that cling to weak connections should be nudged off them.
  • One consistent guest SSID. Every additional network name is another chance for a device to choose badly.

The supporting choices guests never see

  • Resilient local services. DHCP and DNS should be hosted on board with redundancy, so a satellite outage does not also break name resolution.
  • Fair sharing. Per-device limits and uplink QoS stop one large download from degrading everyone else's experience.
  • Redundant core and power. Dual core switches, adequate PoE budget and UPS-backed racks mean a generator changeover does not reboot the network.
  • Low-friction log-on. Captive portals that time out or reappear after roaming are a frequent cause of ‘dropping’ complaints. Simple keys or long-lived sessions usually serve guests better.
  • Local compute and storage. Onboard ruggedised edge computing and storage can host media, caching and management tools, so less depends on the satellite link.
  • Monitoring and remote support. Visibility of link health and client experience means problems are found before the principal notices them.

Questions to ask at new-build or refit stage

  • How is traffic segmented, and what can each segment reach?
  • What happens to a guest video call when the primary uplink fails?
  • How is link health measured, and how quickly does failover occur?
  • Has coverage been validated in the finished interior, including exterior decks?
  • Who monitors the network when the yacht is remote?
  • Is the design documented well enough for the next ETO to understand?

If the answers are vague, the hardware list is not the place to start.

How we approach it

At OceanEdge we design the network first and select hardware to suit it. Segmentation, routing policy, uplink strategy and roaming are defined before a single access point is specified. Our maritime network design and installation work runs from survey to commissioning. Our satellite connectivity services bring VSAT, Starlink Maritime, cellular and shore links together into one managed uplink strategy. For a new build, a refit or an independent review of an existing network, speak to our team.

Frequently asked

Can superyacht guest Wi-Fi ever be completely uninterrupted?
No single link at sea is uninterrupted, but the guest experience can be. With bonded multi-WAN connectivity, redundant core infrastructure and well-engineered roaming, individual link failures pass unnoticed. Guests generally only see an interruption when every uplink is unavailable at once, which is far less common.
Is Starlink enough on its own, or do we still need VSAT?
It depends on itinerary, operational requirements and tolerance for risk. Starlink Maritime offers high throughput and low latency, but obstruction and reliance on a single service are real considerations. Many yachts keep VSAT or cellular as a diverse second path, and the right mix is best decided against actual cruising areas.
What is the difference between multi-WAN failover and bonding?
Failover switches traffic to a backup link when the primary fails, which normally changes the public IP address and breaks active sessions. Bonding tunnels traffic over all available links to a shore-side endpoint with a stable IP. Sessions such as video calls and VPNs can therefore survive a link failure.
Why do guest and crew networks need separate VLANs?
Separation protects onboard systems from guest devices, keeps guest and crew traffic private from each other, and allows independent bandwidth policies. It also contains faults, so one misbehaving device cannot affect the whole vessel. A clearly segmented network is also easier to justify in cyber risk assessments.
Can an existing yacht network be improved without replacing all the hardware?
Often, yes. Many problems come from configuration rather than equipment, such as flat networks, basic failover, poorly tuned roaming or intrusive captive portals. A survey and design review shows what can be fixed in configuration and what genuinely needs replacing, and our team can advise.
OceanEdge Systems
Bespoke AV/IT, connectivity and control for superyachts and megayachts — designed, installed and supported worldwide.