Enterprise Wi-Fi density: the attenuation realities that defeat uniform AP-class deployments

Enterprise Wi-Fi has the most misunderstood unit-of-engineering in the building stack. Tendered by AP count, executed by AP class, audited by signal strength — when the actual metric is client throughput at the worst-case device under the worst-case attenuation profile. Uniform AP-class deployment is the procurement default; mixed AP-class deployment is the engineering answer. The difference shows up in the capex line, the day-two operational reality and the year-three capacity-plan refresh.

## The attenuation envelope is the real constraint

A Wi-Fi 6 access point published at 200 mW with a 24 dBi antenna pattern delivers full-throughput coverage in a free-air rectangular room. Place the same AP in a real building and the attenuation envelope collapses: drywall partitions cost ~3 dB each, brick walls 8-12 dB, reinforced concrete 15-25 dB, low-E glass partitions 5-8 dB, modern office furniture and steel-frame partitions 2-4 dB each. By the time the signal reaches the third room from the AP, the published 1.2 Gbps PHY rate is delivering 60 Mbps to a Wi-Fi 5 client and 200 Mbps to a Wi-Fi 6 client.

## Uniform Wi-Fi 6 is the marketing answer, not the engineering answer

The default enterprise procurement pattern is 'uniform Wi-Fi 6 everywhere' on the basis that newer is better. For a typical three-floor hospitality or commercial building, uniform Wi-Fi 6 deployment costs 2.4× the mixed-class design and delivers no measurable throughput benefit in guest-room corridors and back-of-house — where the per-client throughput is bounded by the user's device class and the application demand, not by the AP class. The cost penalty buys nothing operationally; the PoE budget escalates across the L2 switch stack; the AP-count grows without throughput gain.

## Uniform Wi-Fi 5 is the value answer that collapses at peak

The opposite default — 'uniform Wi-Fi 5, save the capex' — collapses in the high-density zones at peak occupancy. Conference halls at 250 concurrent clients, lobby with 80 simultaneous BYOD devices, dining at 120 concurrent guests — Wi-Fi 5 hits the per-AP throughput ceiling and the 5 GHz channel-reuse wall, with co-channel interference rendering the high-density experience unusable. The capex saving on day one buys an operational headache on every weekend evening.

## Mixed AP-class is the engineered answer

The right engineering answer is mixed AP-class: Wi-Fi 6 four-stream APs in high-density public areas (lobby, conference, dining, atrium), Wi-Fi 5 or entry-tier Wi-Fi 6 APs in lower-density guest zones (corridors, suites, back-of-house, plant rooms). The mesh backhaul is on the managed L2 backbone (Cisco C1300 class, Aruba 6300 class) with per-floor channel assignment that avoids co-channel interference across floors. AP placement is engineered against the attenuation envelope, not against a flat-ground density rule of thumb.

## Predictive survey before procurement

Predictive Wi-Fi survey software (Ekahau, NetSpot, Aruba AirWave, iBwave) consumes the architectural floor plan with material annotations and produces an AP-placement map that holds against the actual attenuation envelope. We run the predictive survey at design stage on every enterprise Wi-Fi deployment above 12 APs; below that, the deployment fits inside the AP catalogue's nominal coverage radius and the survey is informational rather than directive.

## Validation walk after deployment

After the AP plant is installed and configured, the validation walk measures actual signal strength, throughput and roaming behaviour against the predictive map. Client-side throughput is measured at every AP boundary, every elevator lobby, every stairwell entry, every guest-room midpoint. Misplaced APs are relocated against the validation result, not the predictive map. The validation walk is part of the commissioning deliverable, not an optional follow-up.

## Channel plan and co-channel discipline

Per-floor channel assignment on the 5 GHz band avoids co-channel interference between APs on adjacent floors — Floor 2 on channels 36/52/100, Floor 1 on 40/64/108, Ground on 44/60/116. The plan is engineered for the building's actual three-dimensional geometry, not a per-floor independent plan. 6 GHz (Wi-Fi 6E / Wi-Fi 7) opens the channel pool wide enough that co-channel pressure relaxes; deployments that mix Wi-Fi 6E / 7 with Wi-Fi 6 / 5 carry both 5 GHz and 6 GHz plans.

## PoE budget per L2 switch stack

The PoE budget is the per-floor constraint on AP count. A 24-port PoE++ switch with a 740 W budget at IEEE 802.3bt (Wi-Fi 6E / Wi-Fi 7 class) supports 16-20 APs at 30-45 W per AP, not 24. The PoE budget is calculated per stack with 30% headroom for generational refresh; under-budgeting drives port power-cycling under load and is the leading silent failure mode on day-two.

## Callout — what enterprise procurement most miss

**AP count is a procurement metric; per-client throughput at peak occupancy is the engineering metric.** The mixed AP-class deployment delivers protocol-grade roam at a quarter of the capex of uniform Wi-Fi 6, with measurable throughput in every zone. Specify the design against the use case, not the catalogue.

## Reference deployment context

Taraghar State Guest House runs 8 Aruba AP25 (Wi-Fi 6 four-stream) in public areas + 17 Aruba AP11 (Wi-Fi 5) in guest zones across three floors on a Cisco C1300-24T-4G backbone with C921-4P router. Seamless roam across all three floors; protocol-grade reliability for visiting delegations and press; no dead zones at validation walk. Mixed AP-class beats uniform Wi-Fi 6 on capex and uniform Wi-Fi 5 on capacity for this venue class.

## References

1. IEEE 802.11ax (Wi-Fi 6) and 802.11be (Wi-Fi 7) standards.

2. Aruba Instant On / Cisco Meraki / Juniper Mist enterprise Wi-Fi deployment guides.

3. iBwave / Ekahau / NetSpot predictive-survey software documentation.

4. IEEE 802.3bt PoE++ standard — per-port power negotiation.