We test enterprise access points, WLAN controllers, ISP gateways, mesh systems and Wi-Fi clients on the CloudLAB Wi-Fi bench: controlled clients and access points up to Wi-Fi 6, Wi-Fi 7 in progress, attenuation for rate versus range, roaming between APs, and a wired side at 1G and 10G. Throughput per client count, airtime fairness, roam time per handover, latency under load for voice and video, WPA3 and 802.1X scale on the controller.
When the number is needed
A datasheet gives the PHY rate of one client next to the AP in a quiet room. Each moment below asks what the AP does with your client mix, your SSIDs and your floor plan, and ends with a number the team can act on.
Every candidate serves the same clients at the same attenuation steps with the same SSID, security and QoS settings. Vendors see the plan before the first frame and read the same table afterwards.
Ranked table per KPI, client limit per model, 2–3 weeks for 3 APs
The delivered APs and controller with the production config: SSIDs, WPA3 and 802.1X, band steering, QoS, VLANs per SSID, roaming settings. Measured against your design with your client mix, including the cheap ones.
Signed report with pass and fail per case, config exports, from 2 weeks
The acceptance cases run again on the new firmware, the new channel plan or the new roaming profile. What moved, by how much, and whether the cell still clears the design. Production stays untouched.
Diff per KPI against the last run, 3 business days
What we test
The bench is the same, the clients and the pass criteria are not. The table is the starting point of the plan; your floor plan, client mix and applications set the thresholds.
| Device class | Numbers in the report | Clients and config on the bench | Typical size |
|---|---|---|---|
| 01Enterprise access point | Aggregate and per-client throughput, airtime fairness, rate versus range, latency, band steering | Controlled Wi-Fi 6 clients, your SSIDs, WPA3 and 802.1X, programmable attenuation to the AP | Dual- or tri-band AP, 2.5G or 10G uplink |
| 02WLAN controller, on-premises or cloud | APs joined per minute, 802.1X authentications per second, roaming, controller failover under load | Emulated APs and clients sized in the plan, real APs for roaming, your RADIUS and config export | 100–1,000 APs |
| 03Home and ISP gateway | Throughput per band, rate versus range, latency with voice and video running, a 24 h soak | Household client mix, WAN side at 1G to 10G, ISP firmware and defaults, attenuation to the far room | Wi-Fi 5 or Wi-Fi 6 gateway, 1G to 10G WAN |
| 04Mesh system | Throughput behind one to three hops, backhaul rate per hop, steering, recovery when a node drops | Nodes at controlled attenuation, clients on each, backhaul as sold, one node powered off | 2–4 nodes |
| 05Wi-Fi client and IoT device | Throughput and latency against a reference AP, roam time per handover, power save, 802.11k/v | Device next to controlled clients on one AP, attenuation sweep, reference APs, air and wire captures | Laptops, handhelds, cameras, sensors |
Any vendor, enterprise or consumer. Wi-Fi 6 today, Wi-Fi 7 when the bench is ready. TR-398 in-premises cases run on request. Named test tools are picked for the setup and listed in the report.
Around the AP
A throughput number only means something with the right things around the device: the client mix it will serve, the attenuation it will face at the cell edge, and the roams it has to carry without dropping a call
Count and mix from your inventory: 1-, 2- and 3-stream devices, Wi-Fi 5 laptops next to Wi-Fi 6 phones, IoT on 2.4 GHz. The bench emulates them, so the AP sees its own population, not a lab full of flagship phones.
SSIDs, WPA3 and 802.1X with your RADIUS, band steering, QoS and VLAN per SSID, 802.11r/k/v settings, channel width and power from your RF profile. The report shows which setting each loss or slow roam traces back to.
Attenuation steps to the cell edge, a roam to the next AP, an AP rebooted, a controller failed over. Each one with voice and video streams running, timed from the last frame to the first frame back.
What you take away
The report is built for the meeting after the test: the vendor, the integrator and your network team read the same table, with the pass band from the design next to the measured value
Each number comes with its evidence: per-client statistics from the bench, AP and controller counters, air and wire captures and the config export of the run
| KPI | Pass band | Measured | Result |
|---|---|---|---|
| Aggregate at the design client count, near the AP | ≥ 800 Mbps | 870 Mbps | Pass |
| Aggregate at the design client count, WPA3 and 802.1X on | ≥ 800 Mbps | 830 Mbps | Pass |
| Roam time, 802.11r, voice running | ≤ 100 ms | 60 ms | Pass |
| Latency at the design client count, video stream | ≤ 30 ms | 18 ms | Pass |
| 802.1X authentications on the controller | ≥ 50 / s | 72 / s | Pass |
| Band steering, 5 GHz capable clients on 5 GHz | ≥ 95% | 91% | Pass with a note |
Where it runs
APs and controllers ship to CloudLAB and sit on the Wi-Fi bench with controlled clients, attenuators and a wired side at 1G and 10G; cloud controllers are joined from the bench. You watch the run through the portal, with an engineer on every session. The bench serves Wi-Fi 6 today; Wi-Fi 7 is in progress.
How a load run goes, step by step Interoperability: clients, APs and controllers together Routers and switches behind the APs CloudLAB: the bench and how to reach it
Before the brief
Enterprise and consumer access points, on-premises and cloud WLAN controllers, ISP gateways, mesh systems, and clients from laptops to IoT sensors. Any vendor. The bench serves Wi-Fi 6 today, with 2.4 and 5 GHz; Wi-Fi 7 is in progress.
Not yet. The Wi-Fi 7 bench is in progress, and we say so rather than quote a number we cannot measure. Until it is ready, a Wi-Fi 7 AP is tested with Wi-Fi 6 clients, which is what most of its users will be for the next years, and the plan says so on page one.
The count your deployment needs, written into the plan with the device mix and checked on the bench before the run. Per-client attenuation builds the cell edge on purpose rather than finding it by walking; a multi-AP floor is agreed per project.
Yes. SSIDs, WPA3 and 802.1X against your RADIUS or an emulated one, band steering, QoS, VLANs, 802.11r/k/v and the channel plan come from an anonymised export. A vendor-default run can sit next to it to show what the config costs or gains.
TR-398 in-premises cases, rate versus range, multi-client and spatial consistency, run on request on the same bench. Roaming is timed per handover with voice and video streams running, per client type and per profile, between real APs on the bench.
Test plan and quote in 1–3 business days after the brief. A run takes from 2 weeks for one AP model with its controller and 2–3 weeks for a shortlist of 3 models. The report carries every case with pass band and measured value, per-client statistics, captures and config exports; a diff against the last run comes with a re-test.
No prices on the site: scope first, then the quote. The vendor and the integrator can sit in on the brief.
Send the shortlist
Name the models and the KPIs that matter. An engineer replies within one business day with a test plan draft and a quote.
Prefer to chat? Message an engineer on WhatsApp
We use what you send only to answer you. See our Privacy Policy.