Access Point Design

AP placement and the radio environment

Wi-Fi performance depends on where the access point is placed as well as on the radio parameters. An Access Point (AP) broadcasts the radio waves carrying data. A router in a basement corner behind concrete walls and metal pipes leaves areas with a weak signal.

Access point design combines radio-wave physics with the building environment. It covers the location, configuration, and number of APs required by the applications and active clients. The analysis has three areas: Coverage, Capacity, and Channel Planning. Each one affects speed, latency, and the ability to keep a connection.

An indoor AP with an omnidirectional antenna is usually mounted near the ceiling, above obstructions, and close to the middle of the area served by the radio. Cisco documentation notes that wall mounting can limit 360-degree coverage because the wall blocks part of the radiation. The location therefore has to be evaluated together with the antenna type and the intended direction of coverage.

Pillar 1: Coverage planning and dead zones

The goal of AP design is to provide the required signal level where wireless devices operate. A Wi-Fi signal is a radio wave that weakens with distance, reflects from surfaces, and is attenuated by building materials.

Coverage and throughput are tied to the PHY rate. Lower rates use simpler modulation and can work at a weaker signal, but they occupy the channel longer. A design must therefore be checked at the required minimum rate rather than at the highest value listed for the AP.

Placing a single access point

For a small apartment or a single-story home, one well-placed router with an AP can suffice after coverage and capacity are checked. A location near the middle of the used area limits the greatest distance to a client.

• Central location: Place the AP near the middle of the served area. The signal then has a similar path to many rooms, and distant points require less link budget.
• Elevation and open space: Place the AP high, for example on a shelf or near the ceiling, away from furniture. Putting it on the floor or behind a couch increases attenuation and changes the antenna pattern.
• Avoid obstacles and interference: Keep the router away from large obstructions, especially conductive and multilayer materials. Attenuation depends on the material, thickness, and angle of incidence:
  • Metal and concrete: They can strongly attenuate or reflect a signal. A refrigerator, reinforced wall, or metal service duct should be checked during measurements.
  • Brick and plaster: They add attenuation, with the value depending on wall thickness and moisture.
  • Glass and wood: They usually attenuate less than concrete, but coated and multilayer glass can change the result.

  Also avoid sources of RF interference. The 2.4 GHz band shares spectrum with microwave ovens, older cordless phones, baby monitors, and Bluetooth devices.

  An antenna should not sit against a large metal surface. Cisco design material recommends spacing on the order of one or two wavelengths because metal changes the field distribution and can create local signal nulls.

Coverage planning and the site survey

In larger or complex environments such as multi-story homes, offices, or warehouses, one AP is rarely sufficient. Network design starts with a site survey, which maps the radio environment to determine the number and placement of APs.

• Predictive site survey:A designer imports floor plans into software such as Ekahau or NetAlly, then adds virtual walls, doors, and furniture with material properties. The propagation simulation creates a "heatmap" of predicted signal strength and estimates the number of APs before hardware is installed.
• On-site survey:The predictive model must be checked in the building. The "AP-on-a-stick" method places a battery-powered access point on a portable tripod at a proposed location, then the designer measures the signal with a Wi-Fi scanner or survey laptop. Repeating the measurement at each location verifies and refines the design.

Prediction describes assumptions, while the measurement checks real propagation. In a warehouse, the survey must include rack height and stock levels because metal and changing inventory can move the cell boundary. Measurements should use the same antenna type and the locations where clients will operate.

A survey commonly records two coverage metrics among others:

1. RSSI (Received Signal Strength Indicator): This is the received radio power expressed in dBm. A value closer to zero means a stronger signal. Cisco voice WLAN guidance uses a target of $-67$ dBm or better, for example $-60$ dBm is stronger than $-70$ dBm.
2. SNR (Signal-to-Noise Ratio): This metric compares the Wi-Fi signal power with radio noise. It is measured in decibels (dB). A higher value means more separation from noise. Cisco voice WLAN guidance gives a target of at least $25$ dB with a -67 dBm signal.

In the same guidance, a -92 dBm noise floor with a -67 dBm signal gives an SNR of 25 dB. A voice WLAN also uses about 20 percent cell overlap on 2.4 GHz between APs on different channels, so a client can change AP before losing usable signal. These are targets for a specific service profile, and other applications may need different thresholds.

The overlap gives a client another AP to discover before the current signal becomes unusable, but it does not force a handoff. The client selects when and where to roam according to vendor-specific logic. 802.11k can provide a neighbor report, while 802.11r can shorten the security transition when both the client and the infrastructure support it.

Pillar 2: Capacity planning and load

A strong signal in every room does not guarantee good performance. The second pillar is capacity planning, which means estimating active devices, their traffic, and the number of APs required to handle that load.

An AP radio has finite throughput and shares airtime among clients. Adding more devices to one AP increases contention for the medium, so each client receives less speed even when its signal level remains high.

The number of devices associated with an AP does not by itself measure load. A phone can remain associated without transmitting, while one active backup can consume more airtime than several devices sending short requests. The design should count concurrent active clients and assign them application requirements.

Capacity planning factors

• User and device density: First estimate the active devices in an area. A lecture hall, conference room, or stadium concentrates hundreds of clients in a small space and requires a different plan from a home.
• Application requirements: Network traffic has different bandwidth and latency requirements. The design should account for the applications used by clients.

  Application Type	Bandwidth Need	Latency Sensitivity	Examples
  Basic Data	Low	Low	Email, web browsing, IoT sensors
  Voice over Wi-Fi (VoWiFi)	Low	Very High	Voice calls
  Video Streaming	High to Very High	Moderate	Netflix 4K/8K, YouTube
  Real-Time Interactive	High	Very High	Online gaming, video conferencing, AR/VR

A designer can create a "capacity-based design" in which the number of APs follows the user count and required throughput per area. If a lecture hall holds $200$ students and one AP is assumed to handle about $40$ active students, the calculation $200 / 40 = 5$ calls for at least 5 APs even if one or two APs could provide basic signal coverage.

The value of 40 active students is an example assumption. It is not a property of every AP because the result depends on client rates, frame sizes, retransmissions, and the required quality of service. Cisco describes similar ratios as design numbers that must be confirmed with measurements and test traffic.

Pillar 3: Channel planning and co-channel interference

After coverage and capacity are defined, the channels must be planned. Wi-Fi spectrum is shared, so APs on the same channel hear each other and share airtime. This Co-Channel Interference (CCI) reduces throughput in multi-AP networks.

The 2.4 GHz band

The 2.4 GHz band offers 11 or 13 channels, depending on the regulatory domain, but most overlap with their neighbors. Such use causes Adjacent Channel Interference (ACI). A typical 20 MHz plan uses three separated channels, 1, 6, and 11, to limit overlap.

A multi-AP plan in the 2.4 GHz band assigns channels 1, 6, and 11 to nearby access points in a repeating pattern. Physical separation between APs using the same channel reduces the chance that they hear each other and limits CCI.

The 5 GHz and 6 GHz bands

The 5 GHz and 6 GHz bands offer more planning options than 2.4 GHz. The number of usable 20 MHz channels in 5 GHz depends on the regulatory domain, DFS, and TPC rules. Cisco's 802.11a guidance lists up to 23 non-overlapping channels in some geographic areas. The 6 GHz band has up to 59 20 MHz channels in the FCC domain, while the European 5945-6425 MHz lower band provides 24. More channels make CCI easier to limit, but they do not remove external interference or channel reuse in large deployments.

Channel width changes the trade-off between the rate of one transmission and the number of available channels. A 40 MHz channel occupies two 20 MHz blocks, and an 80 MHz channel occupies four. In a dense deployment, 20 MHz channels can therefore provide more reuse opportunities and less CCI, although one client receives a lower PHY rate.

Transmit power and cell size

Maximum transmit power is not automatically the right setting for every AP. In a multi-AP environment, oversized cells overlap, increasing CCI and making it harder for clients to select the appropriate AP. The configured EIRP must also remain within the limits of the regulatory domain.

The designer tunes transmit power so the cell edge reaches the required RSSI with little overlap with neighboring APs on the same channel. This limits CCI and can improve the conditions used by a client when selecting a nearer AP, although the client makes the roaming decision. Too little power creates coverage holes, so the setting must be verified by measurement.

The link budget also includes the client transmitter. An AP may transmit at more power than a phone or sensor, but a weak return signal still limits link quality. Power and cell size should therefore be chosen for the weakest clients and the application with the strictest latency requirement.