Hotel network design: separating guests, staff, payments and devices
Separate guest, staff, payment and device networks to prevent single-point failures. Plan coverage per area and define failover rules to protect hotel operations.
Contents
Key takeaways
- On one flat network a guest laptop, a card terminal and a door-lock encoder can reach each other, so a single compromised device becomes a hotel-wide problem.
- Separation is a decision about traffic before it is a decision about equipment: write down which network may reach which system, then buy hardware that can enforce it.
- Coverage is planned per area and per peak hour, not per square meter, because a pool deck in August and an empty meeting room ask different things of the same access point.
- A second internet line earns its cost only when the hotel has decided in advance which functions it must carry and has tested the switch at a quiet hour.
- The handover file, meaning the network map, address plan, device list and administrator accounts held in the hotel's own name, is the difference between owning a network and renting one.
Hotel network design is the practice of giving each purpose its own network: guests, staff, payments and building devices such as door locks, televisions and cameras. Each network reaches only what its job requires, the flows between them are written down, and coverage and capacity are planned per area and per peak hour.
The front desk loses its connection at 23:40. The card terminal stops, the door-lock encoder stops, and the guest Wi-Fi stops with them. They failed together because nobody ever decided they should be apart. This guide shows you how to make that decision: how to split a network by purpose, what each part may talk to, how to plan coverage for rooms, pool and meeting rooms, what a second internet line should carry, and what an installer must hand over before you pay.
What goes wrong when one network carries everything?
Everything on a flat network can, in principle, reach everything else, so one weak device becomes the whole property's problem. The damage is rarely data theft alone; it is the front desk that cannot issue a key and the terminal that cannot take a payment.
Guest devices are the least controlled equipment on site. Personal laptops and phones with unknown update histories sit in the same address range as the desk computers, and nothing about their condition is known to the hotel.
Long-lived building devices are the second weak point. Door-lock encoders, cameras, televisions and energy controls are updated rarely, often keep the factory password they shipped with, and on a flat network they are reachable from a sun lounger. A device nobody has logged into for three years is exactly the device an intruder wants to find.
One misconfigured device can also stall the shared address range for everyone. A guest router plugged into a wall socket, or a cabling loop, takes down the desk as easily as an attack would. A single flat network hides problems too: when everything shares one segment, nobody can tell whether the breakfast rush or a faulty television is consuming the connection.
The business cost is operational, not abstract. A stopped front desk means a queue in the lobby, keys issued by hand, and payments taken on paper or not at all. The answer is not one box in the rack; it is a written decision about what each group of devices needs, and equipment chosen to enforce that decision.
Watch outThe flat-network failure list
Five consequences worth raising in a risk meeting: a guest device reaching staff systems; a camera or television reachable from the guest network; the shared address range exhausted by guest devices; a payment terminal sharing a segment with guest traffic; and no way to tell which group caused a slowdown.
How do you separate a hotel network by purpose?
Split by purpose, not by floor and not by equipment brand: guests, staff, payments and building devices each get their own network, and each is allowed to reach only what its job requires.
Name the groups first: guest (Wi-Fi, in-room streaming), staff (front desk, back office, housekeeping handhelds), payments (card terminals and the restaurant and spa tills), building devices (door locks, cameras, energy management, lifts, televisions) and a management network for the switches and access points themselves. For the vocabulary, see our companion guide to hotel network and Wi-Fi terms, from access point to VLAN.
There are two ways to separate. Physically separate cabling and switches is the first; the second is one physical network divided into virtual segments, called VLANs, where each segment behaves as if it had its own cabling. The second is cheaper to run and depends entirely on correct configuration.
Wireless separation is not the same thing as network separation. Two Wi-Fi names, one for guests and one for staff, can sit on the same access point, but they must land on different networks further upstream, or the split is only a label.
Payment traffic goes only to the endpoints your acquiring bank and payment provider specify. Ask them to confirm the scope in writing, and treat the installer as an implementer of that answer, not its author. The management network is the one most often forgotten: switches, access points, cameras and the recorder are administered from one restricted place, never from the front desk computer.
Finally, guest-facing devices must not be able to route. A smart television, a printer or a guest router that forwards traffic quietly rejoins the networks you just separated.
| Network | Who and what sits on it | May reach | Must never reach |
|---|---|---|---|
| Guest | Guest Wi-Fi, in-room streaming | The internet only | Staff systems, payments, cameras |
| Staff | Front desk, back office, housekeeping handhelds | The hotel system and printers | Payment provider endpoints, the management network |
| Payments | Card terminals, restaurant and spa tills | Payment provider endpoints only | Guest devices, the internet at large |
| Building devices | Door locks and encoders, cameras and recorder, energy controls, televisions | Their own controllers | The internet except where the vendor requires it; guest and staff networks |
| Management | Switches, access points, recorder administration | Device administration only | Guest traffic |
What may talk to what, and who writes it down?
The map is the deliverable: a short table of which network may reach which system, agreed before installation and updated whenever a device is added. Without it, separation exists only in the installer's head.
Start from the flows that must work: front desk to the hotel system, housekeeping handheld to the same system, card terminal to the provider, lock encoder to the lock controller, cameras to the recorder, and every device to time and updates. Then list what must not work, in the same document: guest to staff, guest to payments, guest to cameras, payments to anything except the provider, building devices to the internet unless the vendor requires it.
Write the rule as default deny. Anything not on the list is blocked, and a new device is a change request rather than something plugged into a spare port. This is the difference between a policy and a habit.
Where the hotel system runs changes the map. A browser-based hotel system needs an outbound connection from the desk browser, not inbound access to the desk computer, which removes a whole category of rules. Network separation and user permissions are two different layers, and a hotel needs both: the network decides which machines may talk, the hotel system decides which people may do what inside it, as covered in our PMS user roles and permissions checklist.
Review the map at every change: a new spa till, a new camera, a new restaurant printer. A map that is six months out of date is worse than no map, because people trust it. To see how the systems on these networks relate to each other, read how the core hotel systems fit together.
ChecklistWhat the traffic map must state
- The name and purpose of each network
- The address range of each network
- Every allowed flow, written as source, destination and the reason for it
- Every explicitly blocked flow
- The owner of the map inside the hotel
- The date of the last review and who signed it off
How do you plan coverage and capacity for rooms, pool and meeting rooms?
Plan per area and per peak hour, not per square meter. Each area has a different number of devices, a different density of people and a different job to do, and the same access point serves them differently.
Survey before quoting. Walk the property with a measuring tool at a busy time, note dead spots and wall materials, and note where people actually sit rather than where the plan says they will. Capacity is about concurrent devices and shared airtime, not only the speed of the internet line: a meeting room with a full session and a corridor with four phones are different engineering problems.
Name the high-density areas in the design: meeting rooms, the lobby at check-in, the pool deck, and the restaurant at breakfast. Spend access points there and accept fewer in corridors and storerooms. Outdoor and pool coverage needs its own decisions: cabling routes, weather protection, and the fact that guests stream video while sitting still for long periods.
Wire what does not move: front desk positions, fixed card terminals, cameras, door-lock encoders and any on-site server or gateway. Cable is still the most predictable medium in the building. Then test at the peak, not at eleven on a Tuesday morning: measure at breakfast and at check-in, record the result per area, and repeat the test after opening.
Do not promise a speed per room. Promise a measured result per area, written into the acceptance terms, so the handover has something to check against. For the legal side of guest Wi-Fi in Türkiye, including login pages and record-keeping, see our companion guide on guest Wi-Fi login pages and record-keeping. Confirm your own duties with a legal adviser.
TipFour questions to ask before signing the quote
Which areas were surveyed, and at what hour? How many devices is each area designed for at peak? Which devices will be wired rather than wireless? What measured result per area will the installer accept at handover?
What should a second internet line carry when the first one fails?
A second line is worth its cost only when the hotel has decided in advance which functions it must keep running and has tested the switch. Two lines with no plan simply fail more expensively.
Define the minimum service in one sentence: guests can be checked in and out, card terminals can take payment, keys can be encoded, and the hotel system is reachable in a browser. That sentence is the specification for the second line.
Choose two providers, ideally entering the building by different routes. Two lines from the same provider usually share the same fault, and a second line in the same duct shares the same digger. Failover must be automatic and rehearsed: pull the cable at a quiet hour, time how long the desk is down, and write the number into the handover file so the next test has a baseline.
Decide what is allowed to fail. Guest streaming, in-room televisions and cloud camera recording can wait; payments and key issuing cannot, and the failover rules should say so. A browser-based front desk changes the requirement in a useful way: the desk needs a working path to the internet, not a server in the back office that also needs protecting.
A phone hotspot is a stopgap for one position, not a plan. Decide in advance who may authorize it and which functions it is allowed to carry. Then write the runbook: who is called, in what order, what is switched off first to protect the payment path, and how the hotel tells guests what is happening.
NoteWhat the failover runbook states
The minimum service list; the order in which functions are restored; the named person who declares a failure; the provider escalation numbers; and the date and result of the last test.
Who manages the network, and what must be handed over?
Whether the work is in-house or outsourced, the hotel must hold the documentation, the administrator accounts and the change log in its own name. Without them it rents its network rather than owning it.
Three workable models exist: full outsourcing under a managed contract; in-house staff with an outside installer for cabling; or a split where the provider runs the network day to day and the hotel holds the accounts and the documentation. All three work. What does not work is the fourth model, where the provider holds everything and the hotel holds a phone number.
Accounts belong to the hotel: administrator access on switches, access points and the camera recorder, the account that owns the equipment, and the internet contracts, all in the hotel's name, with a named deputy. The handover file is a condition of payment, not a favor requested afterwards.
Change control is a habit. Every new device goes through the map, and a photograph of a new switch on the wall is not documentation. Agree the exit in advance: what happens if the provider leaves or is replaced, who returns the credentials, and how long the configuration is kept available.
Guest Wi-Fi and camera recordings involve personal data, so retention periods and who may view what are the hotel's decisions, taken with a legal adviser rather than with the installer.
ChecklistThe handover file
- Network map with the traffic rules
- Address plan per network
- Device inventory with location and serial number
- Administrator accounts in the hotel's name
- Failover runbook and last test result
- Support contacts and response times
- Warranty and contract end dates
What to do next
Before the next quote is signed, ask for two documents: the traffic map and the handover file. Separation, coverage, the second line and the management model all become checkable once those two exist, and the conversation with an installer changes from trust to specification. Everything else in this guide, from which network a camera sits on to what the desk does at 23:40, follows from writing those decisions down before the cabling goes in.
Frequently asked questions
Does separating the network mean buying two internet lines?
No. Separation happens inside the building, on your own switches and access points; the internet line is what all the networks eventually share. A second line is a separate decision about availability, and it should be justified by the functions it must keep running, not by the separation itself.
Can a guest on the hotel Wi-Fi reach the front desk computer?
On one flat network, possibly yes, depending on what else is on that network and how it is configured. Once guests sit on their own network with a default-deny rule, the guest device has no path to the desk unless someone deliberately creates one. The check is the traffic map, not a feeling about the equipment.
How many access points does a hotel need?
There is no responsible number without a survey. The count follows the areas, the number of devices expected at peak, the wall materials and whether the space is indoor or outdoor. Any installer who quotes a count before walking the property is guessing, and the acceptance test should be a measured result per area.
Is a VLAN the same thing as a separate network?
A VLAN is one way to build separate networks on shared cabling and switches: each segment behaves as if it had its own wiring. It is cheaper than running new cable, and it depends entirely on correct configuration and on the rules between segments. Physically separate cabling is the other route, and the two are often mixed.
Who should hold the administrator passwords, the hotel or the IT provider?
The hotel, always, with a named deputy and a sealed record of where they are kept. The provider can hold working access for support, but ownership of the accounts is what lets a hotel change provider without rebuilding its network. This belongs in the contract, not in a conversation after a failure.