TP-Link does not publish a number that says “this handles forty guests.” Nobody does. TP-Link’s business range is really two different products wearing one brand name — a standalone access point with nothing else to buy, and a full Omada ecosystem of gateway, controller and access points — and which one you want changes everything else on this page.
A standalone EAP access point is configured on its own web page, with no controller and no separate gateway to buy — wire it into whatever router is already on site. It talks to us directly. This is the cheaper, simpler path, and it is what most single-room cafés actually want.
The full Omada ecosystem is a gateway (the ER605 or ER7206), an Omada controller, and one or more access points, all managed from one dashboard. Here, the controller — not the gateway and not the access point — is what asks us about a guest, so it is the controller that has to be reachable by us, and the gateway on its own does not include one: an ER605 or ER7206 without a controller cannot run our external sign-in page at all. Omada earns its place once you are running several access points and want them managed together, or want the gateway’s own firewall and VPN features.
No controller, no gateway to buy — these run from their own web page.
Ceiling-mount, standard PoE. The straightforward pick for one room.
More radios, a 2.5G port, needs PoE+. The step up for a busier single room.
Worth knowing: once an EAP is claimed into a controller, its own standalone web page stops being reachable — decide which path you want before you commission it, rather than switching later.
The gateway routes and firewalls the site; the controller is the separate piece that runs the dashboard and is what we actually talk to. Buying just the gateway is not enough on its own.
| Component | Model | What it’s for |
|---|---|---|
| Gateway | ER605 | Smaller site, lower routing throughput |
| ER7206 | Busier site, meaningfully higher routing throughput | |
| Controller | OC200 | Smaller Omada deployments |
| OC300 | Larger Omada deployments, more managed devices |
Worth knowing: a software controller (running on a small server you already have) or TP-Link’s cloud-based controller do the same job as the OC200/OC300 hardware box, at no extra hardware cost, if you have somewhere to run one. The choice between them is about where you want that software running, not about capability.
For a single room, either path copes fine with the internet line most small venues actually have. What actually decides the outcome is whether the room is fully covered, whether your switch or injector can deliver the PoE class the access point needs, and — if you go the Omada route — making sure the controller, not just the gateway, is part of the order.
The other half of sizing is the internet line itself, which is the same arithmetic regardless of which path you choose. We work through it properly, with a worked example, in how much Wi-Fi does a venue actually need — the same maths applies to a restaurant, just at a smaller scale than the hotel example there.
On a standalone EAP, checkout stops a guest signing in again; the session they already hold runs to its own time limit rather than ending immediately.
On the Omada controller path, a genuine cut-off exists on recent controller versions, but it is new enough that it is worth proving with a real phone at commissioning rather than assuming it from the version number. See the TP-Link Omada and TP-Link EAP (standalone) setup pages for the detail on each.
Ranges, not exact figures — nobody, TP-Link included, can tell you a precise number without knowing what your guests actually do. These follow the same method as our hotel guide: a per-guest speed cap, enforced by the equipment, and a contention ratio that assumes not everyone is pulling hard at once. A café or restaurant runs busier concurrently than a hotel room overnight, so the ratio here is tighter than the hotel example.
| Simultaneous guests | Minimum | Preferred |
|---|---|---|
| Up to 10 | 30 Mbps | 50 Mbps |
| Up to 20 | 50 Mbps | 100 Mbps |
| Up to 30 | 75 Mbps | 100–150 Mbps |
| Up to 40 | 100 Mbps | 150 Mbps |
| Up to 50 | 150 Mbps | 200 Mbps |
| Up to 60 | 200 Mbps | 250–300 Mbps |
| Up to 80 | 300 Mbps | 400 Mbps |
| Up to 100 | 400 Mbps | 500 Mbps |
This is independent of which path you pick. The line is what you’re provisioning here, not the box broadcasting it — the ER7206’s own routing throughput is well beyond anything a café-sized line needs. Past 100 simultaneous guests, or once you’re past a single room, this is exactly the point to get a site survey done rather than extrapolate the table further.
No controller, no separate gateway — wire it into whatever router is already on site.
PoE+ needed — check your switch or injector supports it before ordering.
The smaller Omada gateway and controller pairing, for a site that wants central management without over-buying.
Higher routing throughput and a controller built for more managed devices — but AP count and placement should come from the survey, not from this table.
For routing and internet, yes. For our sign-in page, no — the gateway needs an Omada controller (hardware, software or cloud-based) alongside it. The controller, not the gateway, is what talks to us.
Usually not. A standalone EAP talks to us directly, with no gateway or controller to buy, and is the simpler and cheaper choice for one room. Omada earns its place once you want several access points managed together, or the gateway’s own firewall and VPN features.
TP-Link does not publish a practical figure, and neither should anyone else — the rated “concurrent clients” number is a ceiling. Coverage, PoE budget and the internet line matter far more in practice.
On a standalone EAP, no — the session runs to its own time limit. On the Omada controller path, recent versions can, but it is new enough to be worth proving with a real phone at commissioning rather than assumed.
OC200 for a smaller Omada deployment, OC300 once you are managing more devices centrally. A software controller running on a small server you already have does the same job at no extra hardware cost.
Not for a single room — the guidance above is enough. Anything bigger, and walls, materials and where people actually sit decide the access point count and placement far more than which model you buy. Get that surveyed properly before ordering hardware.
Tell us the room, the line speed and how many access points you are planning, and we will tell you which of these fits — and quote for the hardware if you want us to supply it.