// ref · the synthesis principle

Asymmetric Link Visibility

One link, two different views. The client measures the link one way. The AP measures the same link a different way. When those two views disagree, you get a class of field problems that are almost impossible to diagnose from one side alone. This page is the principle that ties together five separate phenomena most engineers learn as unrelated tricks.

Shankar K. · Wi-Fi engineer, 15 years · Last reviewed: June 2026

// the principle

The client's local view of a Wi-Fi link is not the same as the AP's view of that same link. Signal strength, data rate, and even basic reachability can be healthy in one direction and broken in the other. Most field problems that look impossible live in the gap between those two views, and they only become visible when you measure from the right vantage point.

// why the gap exists at all

The RF channel itself is reciprocal. The path loss from A to B is the same as the path loss from B to A at a given instant. So where does the asymmetry come from? Not the channel. It comes from everything attached to the channel: the transmitters run at different power, the receivers apply different thresholds, the two ends run independent rate adaptation, and whoever is holding the capture is standing in a third place entirely. The physics is symmetric. The system bolted onto it is not.

// five instances of one principle
01 Rate adaptation runs independently each way

Uplink and downlink each run their own rate adaptation algorithm. The AP picks the rate it sends to the client. The client picks the rate it sends to the AP. They do not coordinate. A client can hold a high downlink rate while its uplink collapses, or recover on one direction while the other stays stuck. Reading only the downlink rate tells you nothing reliable about the uplink, and a capture taken at one point sees only what reaches that point.

Deep dive: /rssi-mcs-link-budget for the rate, SNR, and link-budget mechanics
02 The sniffer sees its own view, not the AP's

Every RSSI value in a capture is the signal strength as measured by the capture radio, sitting wherever it happens to sit. It is not what the AP heard, and not what the client heard. Put the sniffer near the client and it reads the client's transmissions strong while the AP, further away, hears them weak. Analysts who treat sniffer RSSI as ground truth for the AP's reception make confident wrong calls. The capture point is a third vantage, and you must account for where it stands relative to both ends of the link you are judging.

Capture discipline: /wifi-packet-capture-fundamentals and /wlanpi-pcap-guide
03 The transmitters are not equal

An AP commonly radiates around 20 dBm at 2.4 GHz and 23 dBm at 5 GHz. A laptop sits near 15 dBm. A phone or IoT device runs lower still to save battery and space. The path loss is identical in both directions, but the power launched into it is not, so the downlink arrives strong while the uplink arrives weak. This single mismatch is the most common reason a client shows good signal yet posts a low uplink rate, and it is routinely misdiagnosed as a coverage problem.

AP
~20 dBm (2.4) / ~23 dBm (5)
Laptop
~15 dBm
Phone / IoT
lower, battery limited
Worth knowing: this asymmetry is larger in Wi-Fi than in cellular. A cellular base station divides its power among many users served at once, while a Wi-Fi AP can put full power behind a single downlink. The result is a downlink that can reach much further than the uplink can answer, an imbalance cellular networks largely avoid by design.
04 The AP can choose not to listen

RX-SOP is a vendor knob that sets the signal level below which the AP will not bother to demodulate a frame. Raise it and the AP stops hearing weaker clients, even ones that hear the AP perfectly. The client keeps receiving strong beacons on the downlink, concludes the link is fine, and never roams, while its uplink frames fall below the AP's threshold and go unanswered. The link looks healthy from the client and is dead from the AP. This is the asymmetry made deliberate by configuration.

Deep dive: /rx-sop for the mechanism, vendor behaviour, and the black-hole failure
05 Spatial Reuse decides from the wrong end

802.11ax OBSS_PD lets a transmitter ignore a frame from another BSS if it arrives below a raised threshold, then transmit concurrently. The decision is made at the would-be transmitter, based on what it hears from the other BSS's transmitter. But the device that actually suffers the interference is the destination station of the ongoing transmission, which may sit somewhere else entirely. The reuse decision is taken from one vantage while the consequence lands at another. Same principle: judging a link from the wrong end.

Deep dive: /channel-contention for CCA, OBSS_PD, and the BSS Color mechanics
// the diagnostic implication

Every one of these five is the same shape: a link judged from one end behaves differently at the other. That has a direct consequence for how you troubleshoot. AP-side telemetry alone cannot reveal an uplink the AP is not hearing. Client-side status alone cannot reveal a downlink decision the client never sees. The only way to see the gap is to measure both halves, which usually means a sniffer capture taken with deliberate attention to where it sits relative to the client and the AP.

When a problem looks impossible, the question to ask first is not what is wrong with the link, but whose view of the link you are looking at, and whose view you are missing.

// the five at a glance
Instance Whose view differs Looks like
Rate adaptation Uplink vs downlink rate engine One direction fast, the other stuck
Sniffer vantage Capture radio vs AP vs client RSSI in the capture misread as the AP's
TX power AP power vs client power Good signal, weak uplink rate
RX-SOP AP receive threshold vs client Connected, in range, no traffic, no roam
OBSS_PD Transmitter view vs destination STA Spatial reuse helps one BSS, hurts another
// governing references
IEEE 802.11-2020 · reciprocal channel, independent per-direction rate selection, CCA signal detect
IEEE 802.11ax-2021 · §26.9.2 OBSS_PD-based Spatial Reuse and its transmit-side decision
Transmit power: AP ~20 dBm (2.4 GHz) / ~23 dBm (5 GHz) EIRP, laptop ~15 dBm, phone lower, per vendor and regulatory norms
Related on this site: /rssi-mcs-link-budget · /rx-sop · /channel-contention · /wifi-packet-capture-fundamentals
// share this page
// also on this site
mac
RX-SOP (Receiver Start of Packet)
mac
CCI vs CCC
pcap
Packet Capture Fundamentals
← previous
mac
RX-SOP
next →
rf
RSSI, SNR, MCS & Link Budget
SK
Shankar K., Wi-Fi engineer, Irving TX
Building WiFi Analyser V2 · CWNA-109 in progress · one post every two weeks
// leave a comment