Tutorial 4: IEEE 802.11 Wireless LAN Architecture and Operation (Extended)

BSS = Basic Service Set. The two types are Infrastructure BSS (with AP) and Independent BSS (IBSS, ad‑hoc).

Physical carrier sensing uses CCA to detect energy on the channel. Virtual carrier sensing uses the NAV, a timer set by the Duration field in frames (or RTS/CTS) to indicate how long the medium will be busy.

RTS/CTS reserves the medium for a data transmission, mitigating the hidden terminal problem by notifying all stations within range of either the sender or receiver to set their NAV and defer.

Management (Beacon, Association Request), Control (RTS, CTS, ACK), Data (Data frame, A‑MPDU).

The DS interconnects multiple APs in an ESS, allowing stations to roam between BSSs while maintaining network connectivity.

After a collision, the station doubles its Contention Window (CW) up to CW_max and picks a random backoff count from [0, CW]. This reduces the probability of repeated collisions.

SIFS is the shortest inter‑frame space, used for high‑priority frames (ACK, CTS). DIFS is longer and used for data frames before backoff. SIFS < PIFS < DIFS.

802.11ac with 160 MHz, 256‑QAM (8 bits/sym), 5/6 coding, and short GI gives up to 6.93 Gbps (with 8 spatial streams).

OFDMA divides the channel into resource units (RUs) that can be allocated to different stations simultaneously, reducing contention and overhead, and improving spectral efficiency in dense scenarios.

TWT allows stations to schedule wake times with the AP, reducing power consumption by enabling longer sleep periods and also reducing contention by coordinating transmissions.

BSS Coloring assigns a color to each BSS. Stations can detect the color in the PHY header; if the color matches its own BSS and the signal is weak, it may ignore the transmission and transmit simultaneously, improving spatial reuse.

Authentication verifies the identity of the station (e.g., via 802.1X or PSK). Association establishes a logical connection with the AP, allowing data exchange.

The NAV is a timer set by the Duration field in frames (or RTS/CTS). Stations update their NAV and defer transmission until the timer expires, thereby reserving the medium.

Wi‑Fi 7 introduces 320 MHz channel bandwidth, 4096‑QAM, multi‑link operation (MLO), and enhanced MU‑MIMO for higher throughput (up to 46 Gbps).

A station in power‑save mode indicates its sleep state in its frame headers. The AP buffers its frames. The station wakes up at the next beacon and sends a PS‑Poll to retrieve buffered frames.

802.11ac supports up to 8 spatial streams. 802.11ax also supports up to 8 streams, with improved MU‑MIMO.

A‑MPDU aggregates multiple MAC SDUs into a single large frame, reducing the overhead of MAC headers, ACKs, and contention, thereby increasing throughput.

The Duration field indicates the time (in microseconds) that the medium will be busy after the current frame. It is used to set the NAV for virtual carrier sensing.

802.11n (Wi‑Fi 4) uses up to 4 streams and 40 MHz, achieving 600 Mbps. 802.11ac (Wi‑Fi 5) uses up to 8 streams, 80/160 MHz, and higher modulation (256‑QAM), achieving up to 6.93 Gbps.

Spatial reuse allows overlapping BSSs to transmit concurrently if they can tolerate the interference, using BSS coloring and dynamic CCA thresholds to determine when it is safe to transmit.

Beacons are broadcast periodically by APs to announce the BSS parameters. Probe Requests are sent by stations to actively search for APs; APs respond with Probe Responses.

The random backoff and exponential increase of the contention window on collisions provide a probabilistic fair access, as stations that have recently transmitted are more likely to have larger backoff values, allowing others to gain access.

The Sequence Control field contains the sequence number and fragment number, used for duplicate detection and reassembly of fragments.

Maximum MSDU size is 2304 bytes (with some implementations supporting up to 7935 bytes for A‑MSDU). The maximum MPDU size is 2304 bytes plus overhead.

802.11r uses pre‑authentication and caching of PMK (Pairwise Master Key) to reduce the time required for full authentication and association during roaming, achieving handoff times < 50 ms.

Collision probability p = 1 - (1 - τ)^(n-1) = 1 - (0.95)^4 = 1 - 0.8145 = 0.1855 ≈ 18.6%.

Transmission time = 1500×8 / 100e6 = 120 µs. Total access time = DIFS + backoff (average) + data + SIFS + ACK. Assuming backoff average = 5 slots × 20 µs = 100 µs. Total time = 200 (overhead) + 120 = 320 µs (plus backoff 100 µs → 420 µs). Effective throughput = 1500×8 / 420e-6 = 28.57 Mbps. Actual will be lower due to contention and collisions.

Actual throughput = 300 × 0.7 = 210 Mbps (aggregate). With 10 stations, ideal fair share = 21 Mbps per station, but collisions reduce this further.

Without RTS/CTS, collisions waste channel time. With RTS/CTS, overhead: RTS (20×8/54e6 = 2.96 µs) + SIFS + CTS (2.96 µs) + SIFS + data + SIFS + ACK. For 1500 bytes data = 222 µs. Total with RTS/CTS = 2.96+16+2.96+16+222+16+ (ACK) 20×8/54e6=2.96 → ~278 µs. Without RTS/CTS, just data+ACK ~ 222+16+2.96=241 µs. But collisions may occur; RTS/CTS prevents long data collisions, so for large frames it is beneficial.

The backoff decrements only during idle slots. It decremented 3+2=5 slots, so remaining = 7-5 = 2 slots. It needs 2 more idle slots. Total idle slots = 7. The time = 7 × 9 = 63 µs of idle slots, but the busy periods also consume time. The station will count only idle slots; it will wait through the busy periods (4 slots time) without decrementing. So total time = (3+4+2) slots = 9 slots total (including busy) but backoff counts only idle: 5 idle decrements used, 2 remaining, so it will wait an additional 2 idle slots. Total time from start = (3+4+2+2) slots = 11 slots × 9 = 99 µs.

MCS 8 (256‑QAM, 5/6 coding) with 80 MHz and 3 streams gives 234 data subcarriers × 3 streams × 8 bits/sym × 5/6 × (symbol rate) (typically 4 µs symbol + 0.4 GI → 3.6 µs for short GI? Actually, calculation: 802.11ac MCS 8 at 80 MHz, 3 streams: data rate = 234 × 3 × 8 × 5/6 / (symbol time 3.6 µs) ≈ 1300 Mbps. Check: standard says 1300 Mbps for 3 streams, 80 MHz, MCS 8.

With CSMA/CA, 8 stations would contend, causing overhead (backoff, collisions) and possibly sequential transmissions. With OFDMA, they can transmit simultaneously on different RUs, reducing total airtime to the duration of one packet plus preamble. The improvement factor is roughly the number of RUs (8) assuming perfect scheduling, but with overhead for multi‑user transmission.

Downlink MU‑MIMO: AP transmits to multiple stations simultaneously using different spatial streams. Uplink MU‑MIMO: multiple stations transmit to the AP simultaneously. Uplink is more challenging because stations are distributed and need precise synchronisation and power control; also channel estimation is more difficult due to multiple transmitters.

Scanning: active (Probe Request) or passive (Beacon listening). Then Authentication (or re‑authentication) and Reassociation. Latency depends on scan duration, authentication time (RTT to RADIUS server), association time, and security key exchange (4‑way handshake). With 802.11r, the key exchange is optimised.

BSS coloring allows stations in overlapping BSSs to differentiate transmissions. If a station detects a frame with a different color and the signal level is below a threshold, it may ignore the virtual carrier sensing and transmit simultaneously, increasing spatial reuse. Scenario: Two adjacent offices with separate APs; BSS color enables concurrent transmissions without contention.

PHY data rate for 2 streams, 80 MHz, MCS 11 (1024‑QAM, 3/4) is ~2400 Mbps (approx.). With 75% efficiency, throughput ≈ 1800 Mbps. Actual depends on overhead, contention, etc.

Transmission time = 200×8 / 54e6 = 29.6 µs. Total airtime per frame = 29.6 + 100 = 129.6 µs. Each call sends 50 packets/sec (20 ms). Airtime per call = 50 × 129.6 µs = 6.48 ms/s. With 50% capacity (500 ms/s), maximum calls = 500/6.48 ≈ 77 calls. This is optimistic; actual may be lower due to collisions and backoff.

Assumptions: saturation conditions, ideal channel, n stations, no hidden terminals. The model uses a Markov chain for the backoff stage. The probability τ that a station transmits in a slot is solved from equations involving p (collision probability). The throughput is S = (P_s P_tr L) / ((1-P_tr)σ + P_tr P_s T_s + P_tr(1-P_s) T_c). This is widely used to analyse performance.

For small packets, the overhead of RTS/CTS is significant, and the benefit is low. For large packets, the risk of long collisions due to hidden terminals is high, so RTS/CTS improves throughput. The crossover occurs when the time saved by avoiding data collisions exceeds the RTS/CTS overhead. Typically, for packets > 500 bytes, RTS/CTS is recommended.

EDCA introduces 4 access categories (ACs) with different AIFS and CW parameters. Higher priority traffic has shorter AIFS and smaller CW, giving it more frequent access. TXOP (Transmission Opportunity) allows a station to send multiple frames in a burst, reducing overhead.

At low n, throughput increases with n as more stations contribute to channel utilisation. At high n, collisions increase, reducing throughput. The maximum throughput is achieved at an intermediate n where the channel is optimally utilised.

A‑MPDU aggregates multiple MPDUs into a single PHY frame, reducing MAC overhead. A‑MSDU aggregates multiple MSDUs into one MPDU. Trade‑offs: increased latency (last packet must wait), increased error rate (larger frame more susceptible), and need for Block ACK.

PS‑Poll: station wakes at beacon, sends PS‑Poll to retrieve buffered data. U‑APSD: station triggers delivery when it has data to send. TWT: negotiated scheduled wake times. TWT is most energy‑efficient for IoT as it allows long sleep intervals and predictable scheduling.

In dense environments, 802.11ax improves performance through OFDMA (allocating RUs to multiple users) and MU‑MIMO (serving multiple users simultaneously). It also uses BSS coloring to reduce interference. 802.11ac is limited to SU‑MIMO and does not have OFDMA, so it suffers from contention overhead.

Given target throughput per station S, the total throughput is n × S. Using the Bianchi model, we can solve for the maximum n that yields total throughput (aggregate) and ensure each station gets S. The solution involves solving the fixed‑point equations for τ and p.

Block ACK allows a station to acknowledge multiple frames in a single ACK frame. The sender sends a Block ACK Request (BAR) after a burst of frames, and the receiver replies with a Block ACK bitmap. This reduces the number of ACK frames and overhead, improving efficiency for large transfers.

MLO allows a device to operate on multiple frequency bands (2.4, 5, 6 GHz) simultaneously, aggregating bandwidth and providing load balancing. It can also improve latency by using the best link for each packet and enabling fast fallback.

802.11ax significantly improves efficiency in dense environments, but it cannot completely eliminate congestion due to finite spectrum and coexistence with legacy devices. Network densification, spectrum allocation (6 GHz), and technologies like 5G NR‑U (shared spectrum) are needed to scale capacity.

Management frames (Deauthentication, Disassociation) are unencrypted in legacy 802.11, making them vulnerable to spoofing (DoS attacks). 802.11w adds protection to these frames using robust management frame protection, including encryption and integrity checks.