Wi-Fi 8 (802.11bn) vs Wi-Fi 7: What Changed Interview Prep

802.11be (Wi-Fi 7) vs 802.11bn (Wi-Fi 8) — mechanism-level comparison

Sample interview questions

1. An 802.11bn Multi-Link Device negotiates STR across its 5 GHz and 6 GHz links. What physical condition gates STR operation?
   • A. STR requires the two radio links be sufficiently decoupled - by frequency separation plus antenna isolation of around 30 dB - so the TX leakage from one link does not desensitize the RX of the other. ✓
   • B. STR works at any frequency separation because Multi-Link Operation defines a digital-baseband cancellation algorithm that removes self-interference at the MAC layer.
   • C. STR is gated by the 4096-QAM modulation order; STAs that drop to 1024-QAM lose STR capability and revert to Non-STR alternating operation.
   • D. STR requires both links to use the same channel width and the same Modulation and Coding Scheme so the two link symbol clocks stay frame-aligned.

   Option A is correct because STR (Simultaneous Transmit and Receive) on a Multi-Link Device depends on physical-layer self-interference isolation. When one link transmits, the leakage into the co-located receiver on the other link must be far enough below that receiver's noise floor - usually around 30 dB of combined frequency separation plus antenna isolation - or the receiver desenses and STR is disabled, forcing the MLD into Non-STR or EMLSR mode. Option B is incorrect because Multi-Link Operation does not specify a digital-baseband self-interference cancellation algorithm. The isolation must be achieved physically through RF design and antenna placement; no MAC-layer cancellation removes TX-RX leakage between MLD links. Option C is incorrect because STR mode selection is decoupled from modulation order. A station running 1024-QAM with adequate RF isolation can still operate STR; the gate is isolation, not constellation density. Option D is incorrect because STR explicitly supports asymmetric link configurations. The two links can run different channel widths and different MCS values; nothing in the 802.11be or 802.11bn MLO framework requires symbol-clock alignment across links.

2. A 320 MHz 6 GHz Wi-Fi 8 channel hits a radar incumbent on a 40 MHz subband. How does preamble puncturing keep the rest of the channel usable?
   • A. Preamble puncturing reduces the modulation order to QPSK across the punctured subband while the rest of the 320 MHz channel stays at 4096-QAM, sacrificing rate but not bandwidth.
   • B. Preamble puncturing inserts a guard interval extension over the punctured subband, lengthening the cyclic prefix to absorb the radar pulse without dropping resource units.
   • C. Preamble puncturing signals an unused subchannel pattern in the U-SIG and EHT-SIG fields, so the receiver knows which 20 MHz subblocks to skip when demapping resource units across the 320 MHz aggregate. ✓
   • D. Preamble puncturing requires the access point to fall back to 160 MHz operation whenever any incumbent appears on any subband within the 320 MHz aggregate.

   Option C is correct because preamble puncturing in Wi-Fi 7 and Wi-Fi 8 is a PHY signaling mechanism. The transmitter excludes one or more 20 MHz subblocks from the 320 MHz aggregate and encodes the puncturing pattern in the U-SIG (universal signal field) and EHT-SIG (Extremely High Throughput signal field) preambles. The receiver decodes the pattern, skips the punctured subblocks when demapping resource units, and the remaining 240 or 280 MHz of spectrum carries data normally - keeping the channel partly usable instead of vacating the full 320 MHz. Option A is incorrect because puncturing is a bandwidth-domain mechanism, not a modulation-order adjustment. The MCS chosen for a frame applies across the active subblocks; punctured subblocks carry no payload at any modulation. Option B is incorrect because puncturing does not modify the cyclic prefix length. The CP is fixed by the symbol structure; subblock skipping is signaled in the preamble, not absorbed by guard-interval extension. Option D is incorrect because the puncturing feature exists specifically to avoid the all-or-nothing fallback. An incumbent on one 20 MHz subblock punctures that subblock alone; the access point continues operating on the remaining 240 or 280 MHz aggregate instead of dropping to 160 MHz.

3. Wi-Fi 8 inherits 4096-QAM from 802.11be. What SNR does a 4096-QAM link practically need, and why is it rare in real homes?
   • A. Around 12-15 dB SNR. Higher QAM order provides better performance independent of channel conditions, so 4096-QAM works in typical home conditions whenever LDPC coding is enabled.
   • B. Around 22-25 dB SNR. Home access points sit close enough that this is the typical operating point and 4096-QAM is the default modulation for connected devices.
   • C. Around 35-40 dB SNR at the receiver. The bit error rate floor at 12 bits per symbol demands very low EVM, so 4096-QAM is reachable mainly at short range with strong line-of-sight and well-calibrated transmitters. ✓
   • D. Around 50-55 dB SNR. QAM is constant-envelope like PSK so the power amplifier sees no peak-to-average penalty, but below 50 dB SNR the constellation collapses entirely.

   Option C is correct because 4096-QAM packs 12 bits per symbol, so the minimum Euclidean distance between constellation points shrinks by roughly 6 dB versus 1024-QAM and roughly 12 dB versus 256-QAM. The receiver needs around 35-40 dB SNR (with LDPC coding) to hit the BLER targets required for sustained throughput. In a typical home, distance, walls, and multipath drop the SNR well below this floor; even when the path is short, transmitter EVM (around 2.5 percent rms or better) and crystal phase noise become the binding constraints. The result is that 4096-QAM rates are advertised peaks but real links spend most of their time at 1024-QAM or 256-QAM. Option A is incorrect because the claim that higher QAM order provides better performance independent of channel conditions is a common misconception - higher M actually requires higher SNR for the same BER and must be matched to channel quality. 12-15 dB SNR is the operating range for QPSK or 16-QAM; a 4096-QAM symbol decision at that SNR has an unusable error rate. Option B is incorrect because 22-25 dB SNR sustains roughly 256-QAM or 1024-QAM, not 4096-QAM. The link margin to 4096-QAM's 35-40 dB target is at least 10 dB. Option D is incorrect for two reasons. First, QAM is not constant-envelope like PSK; the QAM constellation has varying amplitude and requires a linear PA, where clipping causes spectral regrowth. Second, shipping 802.11be devices have demonstrated 4096-QAM in lab and short-range field tests, so the 50-55 dB framing of the floor is wrong.

4. Coordinated Spatial Reuse (Co-SR) lets neighboring 802.11bn access points transmit on the same channel. What does Co-SR actually coordinate?
   • A. Co-SR has each access point reduce its transmit power so the cross-talk into a neighboring AP's clients stays below a negotiated threshold, while preserving enough link budget to serve its own clients at the desired MCS. ✓
   • B. Co-SR multiplexes access points onto different OFDMA resource units of the same channel, so each AP transmits on a disjoint set of subcarriers and no spatial coordination is needed.
   • C. Co-SR has the access points exchange channel state information and jointly precode a single stream to a single client across multiple APs, equivalent to cellular Coordinated Joint Transmission.
   • D. Co-SR uses a frequency reuse factor of 1 to maximize capacity on the assumption that aggregate capacity rises monotonically when every AP uses the full channel at full power, with no inter-AP coordination needed.

   Option A is correct because Co-SR is a power-control mechanism layered on top of the existing 802.11ax spatial-reuse framework. Neighboring 802.11bn access points exchange information through a backhaul or over-the-air coordination protocol and agree on a transmit-power ceiling for each that keeps interference into the other AP's served clients below a negotiated threshold. The result is concurrent transmissions on the same channel at lower power, rather than serialized transmissions at full power. Option B is incorrect because OFDMA resource unit allocation is per-AP, not coordinated across access points. Co-SR explicitly allows simultaneous transmission on the same resource units; the spatial separation comes from path loss and reduced TX power, not from RU partitioning. Option C is incorrect because joint precoding across access points is Coordinated Joint Transmission (CJT), a different and more demanding feature in the 802.11bn roadmap. Co-SR coordinates TX power; it does not require shared baseband symbol streams. This is also a useful place to flag a related misconception: spatial multiplexing improves reliability through redundancy. It does not - spatial multiplexing trades reliability for rate, and diversity is the mechanism that improves reliability. Co-SR is neither: it is interference-aware power control, not a multiplexing or diversity scheme. Option D is incorrect because the framing inherits the common misconception that "frequency reuse factor of 1 always maximizes capacity." Reuse 1 maximizes bandwidth per cell but also maximizes interference; the optimal reuse depends on interference management capability. Co-SR exists precisely because uncoordinated reuse-1 leaves capacity on the table; it does not assume aggregate capacity rises monotonically with full-power reuse.

5. Wi-Fi 8 adds PHY-level Hybrid ARQ for the first time in 802.11. How does PHY HARQ improve link efficiency over MAC-layer RTS/CTS retransmission?
   • A. PHY HARQ replaces RTS/CTS reservation; with HARQ enabled the access point stops requesting medium reservation and the contention-free period disappears from the WiFi PHY.
   • B. PHY HARQ allows the receiver to soft-combine the log-likelihood ratios from the original transmission with those from a retransmission, recovering the frame at a lower effective SNR than either attempt would need alone - a saving that pure MAC-layer retransmission cannot deliver. ✓
   • C. PHY HARQ is the same as explicit sounding: the receiver uses NDP-based feedback after every failed frame to refresh the channel state, and the transmitter retransmits with the updated steering matrix.
   • D. PHY HARQ extends WiFi MU-MIMO so 802.11ac uplink multi-user transmissions become possible without OFDMA, with the LTE subframe duration of 10 ms borrowed for the HARQ retransmission window.

   Option B is correct because Hybrid ARQ in 802.11bn is a physical-layer retransmission scheme where the receiver retains the soft demodulator outputs (log-likelihood ratios) from a failed frame and combines them with the LLRs from a retransmission. The combined codeword often decodes successfully at an effective SNR several dB below either transmission individually - the same chase-combining or incremental-redundancy gain that LTE and 5G NR have exploited for years. RTS/CTS at the MAC layer reserves the medium but discards the failed payload entirely; HARQ keeps the energy from prior attempts and converts it into coding gain. Option A is incorrect because PHY HARQ does not replace channel reservation. Medium-access mechanisms including RTS/CTS continue to operate; HARQ adds a physical-layer retransmission capability on top of the existing WiFi PHY MAC. Option C is incorrect because it conflates HARQ with an unrelated WiFi PHY mechanism. The claim that "implicit beamforming is the same as explicit sounding" is itself a known misconception - explicit beamforming uses NDP sounding frames, while implicit beamforming infers CSI from received frames via TDD reciprocity. HARQ is neither of these; it is soft-combining of LLRs across retransmission attempts. Option D is incorrect on two counts. First, WiFi MU-MIMO in 802.11ac is downlink-only; uplink multi-user access requires 802.11ax OFDMA trigger frames, not HARQ. Second, the LTE subframe duration is 1 ms (the frame is 10 ms across 10 subframes), so the borrowed-timing framing is wrong, and HARQ retransmission windows in 802.11bn are scheduled at the PHY without inheriting LTE numerology.

6. Wi-Fi 7 (802.11be) raised the maximum number of spatial streams to 16, sounded as a 16-Tx × 16-Rx channel matrix, which Wi-Fi 8 inherits and leans on harder for coordinated multi-AP MIMO. What is the dominant practical overhead that grows when sounding that 16×16 channel?
   • A. The 16x16 sounding doubles the number of channel taps each receiver must store, doubling the FFT memory footprint at every client.
   • B. The 16x16 sounding adds an extra OFDMA resource unit at every preamble symbol, so the access point loses the equivalent of one user worth of capacity for every sounding cycle.
   • C. The 16x16 sounding requires a 16-stream NDP whose long training fields scale linearly with the number of streams, so the sounding airtime per channel sample roughly doubles versus 8x8 - and the feedback frame containing the compressed beamforming report grows proportionally as well. ✓
   • D. The 16x16 sounding triggers the access point to fall back from MU-MIMO to single-user transmission for the duration of the sounding window, halving aggregate throughput on every NDP exchange.

   Option C is correct because spatial-stream sounding in 802.11 uses a Null Data Packet that carries training fields on each transmit stream. The number of long training field symbols (HE-LTF / EHT-LTF) needed to identify the channel from N transmit streams to M receive antennas scales with the larger of the two; going from 8x8 to 16x16 roughly doubles the sounding airtime per channel sample. The follow-up compressed beamforming feedback frame also grows proportionally because the client must report a 16-column right-singular-vector matrix in quantized angle representation. Together, the sounding overhead is the dominant practical cost of the 16x16 upgrade and bounds how often the channel can be re-sounded under mobility. Option A is incorrect because tap-domain channel storage at the receiver is unrelated to the spatial-stream count. The FFT memory footprint scales with subcarrier count, not with the number of spatial streams. Option B is incorrect because sounding preambles are not OFDMA payload symbols; they are PHY preamble fields. Adding an extra resource unit is not what sounding does - it occupies the preamble, not the data portion. Option D is incorrect because the access point does not fall back to single-user transmission during sounding. The sounding exchange is a brief preamble-and-feedback sequence; MU-MIMO scheduling resumes immediately after the steering matrix is computed.

7. Coordinated Beamforming across two 802.11bn access points needs a backhaul CSI exchange. What is the binding latency requirement on that exchange?
   • A. The CSI exchange must complete within a few hundred microseconds - comparable to the channel coherence interval for indoor walking-speed clients - or the steering nulls miss the moved clients and the inter-AP interference cancellation collapses. ✓
   • B. The CSI exchange can run on a once-per-day schedule because access points are static and the channel between them does not change appreciably across hours.
   • C. The CSI exchange can happen over the air on the same Wi-Fi channel as the served clients with the pilot power doesn't affect estimation error assumption, so backhaul latency is irrelevant and adding more pilots simply improves throughput at scale.
   • D. The CSI exchange is bounded by the Ethernet link MTU because the PMI codebook entry equals the optimal SVD eigenvector, so 1 Gbps wired backhaul delivers exact channel information at the rate required for any Wi-Fi client speed.

   Option A is correct because Coordinated Beamforming steering nulls are computed from a snapshot of the cross-AP channel state and they remain valid only as long as that channel is stationary. For walking-speed clients in the 5 GHz and 6 GHz bands, the channel coherence interval is on the order of a few hundred microseconds to a couple of milliseconds. If the CSI exchange across the backhaul plus the steering computation takes longer than that, the nulls are pointed at where the client used to be - and the interference cancellation gain disappears. This is why Co-BF deployments typically depend on wired Ethernet backhaul with deterministic forwarding, not a contended wireless mesh hop. Option B is incorrect because the binding timescale is the wireless channel coherence interval, not the AP placement. The APs are static, but the clients and the multipath environment around them are not. A once-per-day CSI exchange would produce steering nulls aimed nowhere useful. Option C is incorrect on multiple grounds. First, pilot power does affect estimation error - the MMSE estimation MSE is 1/(1 + SNR_pilot), so power-boosted pilots improve accuracy. Second, the claim that more pilots improve throughput is a known misconception: increased pilot density actually reduces spectral efficiency, and the optimal density balances estimation accuracy against overhead. Third, Co-BF deployments need a dedicated backhaul because contention on the served-client channel pushes the exchange latency past the coherence interval. Option D is incorrect because the bound is the wireless coherence interval, not the Ethernet MTU. The framing also inherits the misconception that the PMI codebook entry equals the optimal SVD eigenvector - in practice PMI is a quantized codebook index and chordal distance to the true eigenvector introduces residual error. Limited feedback CSI does not give exact channel information; codebook quantization causes a typical 1-2 dB loss versus perfect CSI on Type-I codebooks.

Frequently asked questions

What is Wi-Fi 8 and how does it relate to 802.11bn?

Wi-Fi 8 is the common name for IEEE 802.11bn, the next Wi-Fi generation after Wi-Fi 7 / 802.11be. The public direction for 802.11bn is Ultra High Reliability: steadier latency, lower packet loss, better roaming, and more predictable performance in dense or interference-limited networks.

How is Wi-Fi 8 different from Wi-Fi 7?

Wi-Fi 7 pushed peak throughput with 320 MHz channels, 4096-QAM, and Multi-Link Operation. Wi-Fi 8 keeps the broad waveform baseline but shifts the emphasis toward reliability, coordination, and real-world stability, especially in crowded deployments and mobility-heavy environments.

What does multi-AP coordination mean in Wi-Fi 8?

It means neighboring APs are expected to coordinate spectrum and spatial decisions instead of acting independently. The goal is to reduce contention and interference by making multiple APs behave more like a coordinated system.

What is the role of MLO in Wi-Fi 8?

Multi-Link Operation first appeared in Wi-Fi 7, but Wi-Fi 8 is expected to refine it for reliability and latency control. In interview terms, the important idea is not only using multiple links for more throughput, but steering traffic across links to reduce tail latency and packet loss.

Will Wi-Fi 8 add a new band or much higher modulation?

The public direction does not point to a new band or a dramatic jump in constellation order. The standard effort is focused on better use of the existing 2.4 GHz, 5 GHz, and 6 GHz ecosystem through coordination, mobility, and reliability improvements.

Why is Wi-Fi 8 useful for interview prep?

Because it shows whether you can explain the next step in Wi-Fi evolution without reducing it to "faster than Wi-Fi 7." Interviewers want to hear about reliability, coordination, roaming, dense deployments, and the tradeoff between peak rate and usable service quality.

What is Non-Primary Channel Access (NPCA) in 802.11bn, and what problem does it solve?

NPCA is an 802.11bn channel-access mechanism that lets an access point or station transmit on a non-primary 20 MHz subchannel when the primary channel is busy. It addresses the "different view" problem — situations where the AP and a station have different views of channel state because of obstacles, attenuation, or distant interferers. NPCA uses a secondary EDCA procedure on the non-primary channel and is gated by an RSSI threshold so it is only enabled when the AP-to-station distance is short enough to bound hidden-node risk. In interview answers, the useful framing is that NPCA expands when transmission is possible without breaking the contention safety properties of CSMA/CA.

What is Coordinated Beamforming (Co-BF) and how does sounding work across access points?

Coordinated Beamforming is a Wi-Fi 8 (802.11bn) mechanism where neighboring access points sound the channel together and form transmit beams that null interference toward each others served clients, instead of each AP beamforming independently. Two sounding sequences are under discussion at the working group: joint sounding, where APs sound simultaneously, and sequential sounding, where they sound one after the other. The trade-off is recovery: sequential sounding lets the initiating AP detect a missing response after SIFS and retry, while joint sounding cannot recover as cleanly if a responder misses its NDP. For interviews, the engineering point is that Co-BF needs a low-latency backhaul between APs because the steering matrix is only valid for the channel coherence interval — a few hundred microseconds at walking-speed clients.

How does Dynamic Sensitivity Control (DSC) interact with NPCA in 802.11bn?

DSC adapts the CCA (Clear Channel Assessment) threshold based on network density and interference — raising it expands spatial reuse but increases hidden-node risk; lowering it shrinks spatial reuse but improves contention safety. NPCA depends on getting that threshold right because non-primary channel transmission must not collide with a far station that the AP cannot hear. In a well-designed 802.11bn deployment, DSC and NPCA share the same threshold view: if DSC is set aggressively for reuse, NPCA gating must tighten correspondingly. In interview terms, this is the kind of cross-mechanism interaction that distinguishes feature memorizers from candidates with a real systems mental model.

Related topics