A marina looks like an easy Wi-Fi environment. Wide open, almost no walls, clear sight lines from the shore across the whole basin. In practice it is one of the more demanding sites you can be asked to cover, and the reasons are a mixture of physics, chemistry and access.
Coverage has to reach devices sitting low in vessels, over a surface that reflects strongly and moves with the tide, through a population of obstructions that arrive and leave on their own schedule, using equipment mounted in salt air on structures that flex — and fed by cables that have to reach positions hundreds of metres from the nearest building.
Propagation over water is not free space
Open water is close to an ideal reflector at the shallow grazing angles typical of a marina link. That produces a strong, coherent reflected path alongside the direct one, and the two arrive at the receiver with a phase difference set by the geometry — antenna heights and horizontal distance.
Where they arrive in phase they add and the signal is better than free space would suggest. Where they arrive out of phase they cancel, and you get a deep null. Over water, the reflected path is strong enough that the cancellation can be severe rather than a mild ripple.
The consequence is a coverage pattern with structure in it: alternating stronger and weaker regions as you move away from the antenna, and — importantly — as you change height. A device on a flybridge and a device in a saloon at waterline are at different points in that pattern.
Multipath has a second effect worth knowing about. MIMO systems get their throughput from being able to separate multiple spatial streams, and that separation depends on the paths being sufficiently different from one another. A link that is dominated by one strong direct path and one strong specular reflection presents a highly correlated channel, and the achievable rate can be lower than the excellent signal strength implies. This is the over-water version of "full bars, poor throughput", and it surprises people who assume line of sight guarantees performance.
Line of sight is necessary, not sufficient
Two further things get in the way of the assumption that if you can see the berth, you can serve it.
Fresnel zone clearance. A radio path is not a pencil line; the energy travels in a set of ellipsoidal zones around it, and the first of those needs to be substantially clear of obstructions for the link to behave like free space. Over water the reflecting surface itself intrudes into that zone, which is precisely why the reflection is so strong. Raising an antenna gains clearance over pontoon-level obstructions but also changes the reflection geometry — so higher is not automatically better, and the height is a design decision rather than a default.
The client is inside a boat. Coverage that reaches the vessel does not necessarily reach the phone in the cabin. A GRP hull and superstructure passes RF tolerably. A steel or aluminium hull is close to a Faraday cage: signal enters through windows, hatches and companionways, and coverage below deck can be poor even with a strong signal on the pontoon. Metallised or coated glazing on modern motor yachts attenuates significantly. Devices low in a hull are also behind the water surface line, shadowed by the hull structure from anything on the shore.
What follows from that is a design that puts signal down the pontoon at low level, close to the vessels, rather than relying on a small number of high-powered positions on the shore trying to cover the whole basin. Long-range shore-mounted coverage is often the right answer for open water, moorings and the approach, and the wrong answer for berths.
Boats are obstructions that move
The RF environment of a marina is reconfigured continuously by its users, and nobody informs the network.
- Large vessels shadow small ones. A substantial motor yacht on an outer pontoon is a metal or composite wall several metres high. Berths behind it lose the path they had yesterday.
- Masts and rigging scatter. A forest of aluminium masts, standing rigging and radar arches is an excellent scattering environment. It adds multipath rather than blocking outright, but it makes the channel unpredictable.
- Occupancy changes the geometry, not just the load. An empty pontoon in February and a full one in July are different propagation environments as well as different capacity problems.
- Berths are not fixed. Visitor berthing moves vessels around; a rafted-up pontoon during an event bears no resemblance to the plan.
The practical response is to design with margin along the pontoons rather than to a knife edge, and to state the conditions a survey was taken in — occupancy, season, state of tide. Those are the variables, and naming them is the difference between a useful report and a misleading one.
Enclosures, corrosion and material choice
Marine atmosphere is one of the harsher environments commercial equipment is asked to survive, and the failures are slow and expensive rather than immediate.
Ingress protection
Outdoor access points and enclosures need an IP rating appropriate to the exposure — typically IP66 or better for positions taking wind-driven spray, and higher where washdown or immersion is credible. The rating alone is not the whole story:
- Glands and entries. Most water gets in through the cable entries, not the housing. Correctly sized glands, entries facing downward, and drip loops so water runs away from rather than into the gland.
- Condensation. A perfectly sealed box in a location that heats in the sun and cools overnight will accumulate moisture internally. Breather or pressure-equalisation vents are how that is managed; sealing harder makes it worse.
- UV. Enclosures, cable sheaths, ties and gaskets all degrade in sunlight. Marine-grade and UV-stable materials throughout, and expect black plastic ties to be a maintenance item.
Corrosion
- Fixings. A316 stainless rather than A2/304 where salt exposure is real. The cost difference is trivial against the cost of returning to a pontoon position to replace a seized bracket.
- Galvanic corrosion. Dissimilar metals in contact in the presence of an electrolyte — and salt spray is an excellent electrolyte — corrode preferentially. Stainless fixings bolted directly to an aluminium enclosure or an aluminium pontoon structure is the classic case. Isolate with suitable washers or bushes.
- Connectors. External RF and Ethernet connections need weatherproofing beyond the connector's own seal: self-amalgamating tape properly applied, then an outer protective wrap. Connectors are the most common single point of marine failure and the easiest to get right.
- Antenna and mount hardware. Marine-grade or anodised aluminium, stainless banding for pole mounts, and no ferrous hardware anywhere it can weep rust across somebody's gelcoat.
Surge and bonding
Exposed masts and pontoon-mounted equipment on open water are prominent positions. Surge protection on Ethernet runs entering buildings, and equipment earthing and bonding coordinated with the marina's existing electrical arrangements, belong in the design. Where a run crosses between structures, fibre has a real advantage here beyond distance: it carries no conductive path between them at all.
Mounting, power and containment to remote positions
This is the part of a marina project that consumes the programme, and it is usually the part that was scoped last.
Where equipment can actually go
- Pontoons put the signal where the boats are, which is the RF answer. They also move, flex under load and wave action, and are often aluminium with limited structure to fix to. Fixings need to tolerate continuous small movement without fatiguing, and service bollards, lighting columns and finger-berth posts are usually better hosts than the deck structure itself.
- Shore buildings and outbuildings — offices, facilities blocks, boatyard sheds, harbourmaster's buildings — give stable mounting, easier power and easier access. They are further from the vessels and higher, which brings the reflection geometry back into play, and they suit sector antennas covering a basin rather than omnidirectional coverage of a berth.
- Lighting columns and existing structures often already have power and a route, which is worth establishing before designing a new one.
- Access for maintenance. A position that requires a boat, a low tide and two people is a position you will visit reluctantly. That should influence where equipment goes, not merely how it is priced.
Getting a cable there
- The 100 metre limit is real. A balanced twisted-pair Ethernet channel is 100 m end to end, conventionally 90 m of fixed cable plus patch leads. Marinas routinely need to reach further than that, and the distance from the head of the pontoon to the far berths alone can exceed it.
- Fibre with a local node is the usual answer: fibre to a small weatherproof enclosure part-way along, containing a media converter or small switch and injecting PoE locally for the access points around it. This also solves the bonding problem between structures.
- That node needs power. Which means a local supply, an enclosure to house it, and thought about what happens when the pontoon supply is isolated for maintenance. Pontoon service ducts frequently already carry power and water to the berth bollards, and using them requires segregation and cooperation with whoever maintains them.
- Ducts flood. External-grade, gel-filled or otherwise water-blocked cable, terminated in enclosures above the highest water level you can credibly expect, not at the first convenient point.
- PoE budget. Outdoor access points draw more than indoor ones, particularly where heaters, multiple radios or additional radio interfaces are involved. Know which standard each position needs — the 802.3af, at and bt classes differ substantially in what they deliver — and budget the switch or injector accordingly rather than discovering it at commissioning.
All of this is physical infrastructure work, and it is where marina projects most often stall when the RF design and the installation are done by different companies. It is one of the clearer arguments for keeping the survey, the design, the containment and the cabling in the same hands.
Seasonal load and network separation
A marina's demand curve is unlike almost any other site. Winter is a small number of liveaboards and staff, using the network heavily and continuously. Summer is a large number of visitors for short periods, with a peak that may be a single regatta weekend. Designing for the average serves nobody.
The workable approach is to design the coverage for the geography — which does not change — and the capacity for the peak, then use policy to manage the difference. Capacity headroom in a marina usually comes from more, smaller cells along the pontoons rather than from higher power, for the same contention reasons that apply anywhere else.
Three networks, not one
- Berth holders. Long sessions, high expectations, and in the case of liveaboards a genuine domestic internet requirement. Often a paid or tiered service, which means identity, session control and fair-use policy.
- Visitors. Short stays, high turnover, minimal onboarding tolerance. Usually a portal or voucher, ideally tied to the berthing system so access expires with the booking.
- Operational. Fuel berth, boatyard, travel hoist, CCTV, access control and gate systems, ANPR, weather stations, and metering for pontoon power and water. These are the systems that must keep working when the guest network is saturated on a bank holiday, and they should be on their own VLANs with their own policy and — where the design allows — their own capacity guarantee.
Separating them is the same discipline as any guest network: distinct VLANs, explicit firewall policy, client isolation on the public networks, and management interfaces unreachable from any of them. What is specific to a marina is that the operational network includes systems with real safety and revenue consequences, sitting outdoors, on the same infrastructure as the visitor Wi-Fi. That is worth designing deliberately. More on marina and waterside work.
Let's find out what your Wi-Fi is actually doing.
Book a survey and you get measured data, a design you can build to, and a number you can budget against — not an opinion.