Wireless Networking Intermediate

Wi-Fi sends data over radio waves — electromagnetic energy that travels through the air. A wave’s frequency is how many cycles pass per second, measured in hertz (Hz); Wi-Fi operates in the gigahertz (GHz, billions of cycles per second) range. Wavelength is the physical distance of one cycle, and it is inversely related to frequency: higher frequency means shorter wavelength.

Frequency matters because it controls a fundamental trade-off. Lower frequencies (longer waves) travel farther and pass through walls more easily; higher frequencies (shorter waves) carry more data but are absorbed by obstacles more quickly, so they cover less distance.

Wi-Fi uses unlicensed bands. The 2.4 GHz band offers the best range and wall penetration but is narrow (few channels) and crowded — microwaves, Bluetooth and cordless phones all share it. The 5 GHz band has far more channels and less interference, giving higher throughput, but its shorter waves cover less distance and are stopped more easily by walls.

The newer 6 GHz band, opened for Wi-Fi 6E and Wi-Fi 7, adds a large block of fresh spectrum with many wide channels and very little legacy congestion, at the cost of even shorter effective range than 5 GHz.

Wi-Fi is defined by the IEEE 802.11 standards, each a letter suffix. 802.11a (5 GHz) and 802.11b (2.4 GHz) came first, followed by 802.11g (2.4 GHz, faster). 802.11n introduced MIMO (multiple antennas) and works on both bands. 802.11ac pushed 5 GHz throughput much higher, and 802.11ax improved efficiency in dense environments, especially with many clients.

To simplify the alphabet soup, the Wi-Fi Alliance gave them generation numbers: 802.11n = Wi-Fi 4, 802.11ac = Wi-Fi 5, 802.11ax = Wi-Fi 6 (and Wi-Fi 6E when it uses the 6 GHz band).

Each generation raised the ceiling. 802.11b topped out at about 11 Mbps; 802.11a/g reached 54 Mbps. 802.11n (Wi-Fi 4) used wider channels and MIMO to reach hundreds of Mbps. 802.11ac (Wi-Fi 5) reached the gigabit range, and 802.11ax (Wi-Fi 6) pushed higher still while focusing on real-world efficiency.

These are theoretical link rates, not actual file-transfer speeds. Real throughput is always lower because of protocol overhead, shared airtime, distance, interference and the fact that Wi-Fi is half-duplex — devices take turns rather than sending and receiving at once.

Each band is divided into channels — specific slices of frequency an access point transmits on. A wider channel carries more data: 2.4 GHz channels are typically 20 MHz wide, while 5 GHz and 6 GHz support 40, 80 and even 160 MHz channels for higher throughput.

But width is a trade-off. A wider channel occupies more spectrum, so fewer non-overlapping channels fit in the band, raising the chance of interference. In congested areas a narrower channel can actually perform better because it competes with fewer neighbours.

The 2.4 GHz band has channels numbered 1–13 (region dependent), but they are spaced only 5 MHz apart while each 20 MHz channel is much wider — so adjacent channels overlap. The classic result is that only channels 1, 6 and 11 are far enough apart not to overlap one another.

Best practice in a multi-AP deployment is to put neighbouring access points on different members of the 1/6/11 set so their signals do not interfere. Choosing channel 3 or 9, by contrast, smears energy across two of the good channels and degrades everyone’s performance.

There are two flavours of channel interference. Co-channel interference (CCI) happens when two access points share the same channel: they politely take turns (they hear each other and share airtime), so nobody’s data is corrupted but capacity is divided and latency rises. Adjacent-channel interference (ACI) happens when APs sit on partially overlapping channels (say 1 and 3): they cannot decode each other, so their signals collide as noise, corrupting frames and forcing retransmissions.

Counter-intuitively, co-channel is usually the lesser evil. Reusing 1/6/11 deliberately creates co-channel (manageable) and avoids the more damaging adjacent-channel overlap.

The SSID (Service Set Identifier) is the human-readable network name you pick from a list, like CoffeeShop-WiFi. The BSSID is the MAC address of a specific access point’s radio — it uniquely identifies one AP even when several broadcast the same name. The ESSID describes an extended service set: many APs sharing one SSID so a client can roam between them as if it were a single large network.

So a shopping mall might advertise one SSID (Mall-Free-WiFi) across dozens of APs (each with a distinct BSSID), all forming one ESS.

In infrastructure mode (the normal Wi-Fi setup) clients connect through a central access point (AP) rather than directly to each other. The AP periodically sends beacon frames announcing its SSID and capabilities. A client discovers networks by listening for beacons (passive scanning) or sending probe requests (active scanning).

To join, the client goes through authentication and then association, after which the AP forwards its traffic to and from the wired network. A client is associated with exactly one AP at a time; to use a different AP it must re-associate.

Wired Ethernet historically used CSMA/CD (Collision Detection): a station sends and, if it detects a collision on the wire, stops. Wireless cannot reliably detect its own collisions while transmitting, so Wi-Fi uses CSMA/CA — Carrier Sense Multiple Access with Collision Avoidance.

With CSMA/CA a device first listens to check the channel is idle, waits a short interval plus a random backoff, and only then transmits. The receiver returns an ACK to confirm delivery; if no ACK arrives, the sender assumes a collision and retries. Because the medium is shared and half-duplex, only one device can transmit successfully at a time.

The hidden-node problem arises when two clients can both reach the AP but cannot hear each other (for example, on opposite sides of a building). Each senses the channel as idle, transmits at the same time, and their frames collide at the AP — even though CSMA/CA worked correctly from each one’s point of view.

The fix is the optional RTS/CTS handshake. A station sends a short Request To Send; the AP replies with a Clear To Send that every client in range hears, reserving the medium for the duration. This silences the hidden node and prevents the collision, at the cost of extra overhead.

An antenna shapes where radio energy goes. An omnidirectional antenna radiates roughly equally in all horizontal directions — good for general coverage, like a home router. A directional antenna (such as a Yagi or panel) focuses energy into a narrow beam, reaching farther in one direction, ideal for point-to-point links between buildings.

Antenna gain measures this focusing and is expressed in dBi (decibels relative to an isotropic radiator). Higher dBi means a more concentrated beam, not more raw power: a high-gain omni effectively flattens the doughnut of coverage, trading vertical reach for horizontal distance.

RSSI (Received Signal Strength Indicator) reports how strong the received Wi-Fi signal is, usually in dBm. These are negative numbers: a value closer to zero is stronger, so -50 dBm is excellent while -80 dBm is weak. Roughly, -67 dBm or better is recommended for reliable applications.

Raw strength is not the whole story. SNR (Signal-to-Noise Ratio) compares the signal level to the background noise floor; a higher SNR means a cleaner link and faster, more reliable rates. A strong signal in a very noisy environment can still perform poorly if the SNR is low.

To cover a large area you deploy several APs that share one SSID (an ESS). As a client moves, its signal to the current AP weakens; when a better AP is available the client roams — it disassociates from the old AP and associates with the new one. The client, not the network, decides when to roam, based on signal thresholds.

Good roaming design overlaps coverage so there is always a strong AP nearby, and uses the same SSID with different channels (and BSSIDs) on neighbouring APs. Standards like 802.11r speed up the re-authentication so roaming does not interrupt calls or video.

Early Wi-Fi used WEP (Wired Equivalent Privacy), which is thoroughly broken — its weak RC4 key scheme can be cracked in minutes, so it must never be used. WPA was a stop-gap that added TKIP to patch WEP-era hardware, but it too is now considered insecure.

WPA2 became the long-time standard, using strong AES encryption via the CCMP protocol. The current best practice is WPA3, which strengthens authentication with SAE (Simultaneous Authentication of Equals, also called Dragonfly), protecting against offline password-guessing and adding forward secrecy.

In WPA2/WPA3 Personal, both the client and the AP already know the shared passphrase (turned into a key called the PMK). They never send the passphrase over the air. Instead they run the 4-way handshake: a four-message exchange that proves each side knows the key and derives fresh session keys (the PTK) for actually encrypting traffic.

The handshake also exchanges random nonces so the session keys are unique each time, and confirms both sides agree before data flows. WPA3’s SAE replaces the vulnerable parts of this process to resist the offline dictionary attacks that affected WPA2.

A guest network is a separate SSID that gives visitors internet access while keeping them away from your private devices. Behind the scenes it is usually mapped to a separate VLAN/subnet so guest traffic cannot reach internal servers, printers or file shares.

Client isolation (also called AP isolation) goes further: it stops devices on the same Wi-Fi from talking to each other, allowing only traffic to the gateway. This is common on guest and public hotspot networks so one infected laptop cannot attack the phone next to it.

Many everyday devices disturb Wi-Fi, especially in 2.4 GHz: microwave ovens, Bluetooth devices, cordless phones, baby monitors and even poorly shielded cabling. Physical obstacles matter too — thick walls, metal, and water (including crowds of people) absorb signal and shrink coverage.

A site survey measures real-world signal and interference before or after deployment. A passive survey listens to existing signals to map coverage and noise; an active survey connects to the network to test throughput and roaming. The results guide AP placement, channel selection and power levels for reliable coverage.