“Wireless Ethernet” is the everyday name for what’s technically the IEEE 802.11 family of standards — the technology behind Wi-Fi. Despite the name, Wi-Fi isn’t literally Ethernet transmitted over the air; rather, it’s designed to be compatible with traditional wired Ethernet at the network layer, while using fundamentally different mechanisms at the physical and link layers to deal with the unique challenges of a shared, wireless medium. This article explains, from first principles, how wireless Ethernet networks actually work — from radio frequencies to frame transmission to the challenges of a shared airwave medium.
Wireless vs. Wired Ethernet: The Fundamental Difference
Wired Ethernet has dedicated physical paths (cables) between devices and a switch, with essentially no possibility of two devices’ signals colliding in the same physical medium (thanks to modern switched, full-duplex Ethernet). Wireless Ethernet has no such luxury — it broadcasts through open air, a medium every device within range shares.
flowchart TD
A[Wired Ethernet] --> B[Dedicated Cable per Device to Switch]
B --> C[No Collision - Full Duplex Switching]
D[Wireless Ethernet - 802.11] --> E[Shared Radio Spectrum]
E --> F[Collision Avoidance Required - CSMA/CA]The Physical Layer: Radio Frequencies and Channels
Wi-Fi operates primarily in these frequency bands:
| Band | Frequency Range | Characteristics |
|---|---|---|
| 2.4 GHz | ~2.4–2.5 GHz | Longer range, more interference (shared with Bluetooth, microwaves), fewer non-overlapping channels |
| 5 GHz | ~5.15–5.85 GHz | Shorter range, less interference, many more non-overlapping channels, higher potential speeds |
| 6 GHz (Wi-Fi 6E/7) | ~5.925–7.125 GHz | Newest band, least congestion, requires newer hardware |
Each band is divided into channels — specific frequency slices that devices tune to, similar to radio stations. In the 2.4 GHz band, channels 1, 6, and 11 are commonly used because they don’t overlap with each other, avoiding interference between nearby networks using different channels.
The Core Challenge: A Shared, Half-Duplex Medium
Unlike switched wired Ethernet, only one device can successfully transmit on a given wireless channel at any given moment without the signals interfering with each other (colliding). This is why Wi-Fi uses a fundamentally different access method than wired Ethernet.
CSMA/CA — Carrier Sense Multiple Access with Collision Avoidance
Wired Ethernet historically used CSMA/CD (Collision Detection) — a device could detect a collision while transmitting and simply retransmit. Wireless can’t reliably detect collisions this way (a transmitting radio can’t simultaneously listen for a collision easily), so Wi-Fi uses CSMA/CA — Collision Avoidance instead:
sequenceDiagram
participant Device
participant Medium
Device->>Medium: Listen - is channel busy?
Medium-->>Device: Channel is idle
Device->>Device: Wait a random backoff time
Device->>Medium: Transmit frame
Medium-->>Device: ACK received (success)- Before transmitting, a device listens to check if the channel is currently in use
- If busy, it waits
- If idle, it waits an additional small random backoff period (to reduce the chance two devices start transmitting at exactly the same instant)
- It transmits its frame
- The receiver sends back an acknowledgment (ACK); if no ACK arrives, the sender assumes a collision or failure occurred and retries
Association: How a Device Joins a Wireless Network
Before a device can send any data, it must go through several steps to join (associate with) an access point:
sequenceDiagram
participant Client
participant AP
Client->>AP: Probe Request (or hears Beacon)
AP-->>Client: Probe Response / Beacon (advertises SSID, capabilities)
Client->>AP: Authentication Request
AP-->>Client: Authentication Response
Client->>AP: Association Request
AP-->>Client: Association Response
Client->>AP: EAPOL 4-Way Handshake (WPA2/WPA3 key exchange)- Discovery: The client either passively listens for beacon frames (which APs broadcast regularly, advertising their SSID and capabilities) or actively sends probe requests
- Authentication: An initial (often just formal, in modern WPA2/WPA3 networks) authentication exchange
- Association: The client formally joins the AP’s service set
- Key exchange (4-way handshake): For WPA2/WPA3 networks, a cryptographic handshake establishes session encryption keys before any real data can flow
Frame Structure: How Wi-Fi Frames Differ From Wired Ethernet Frames
While both ultimately carry the same higher-layer data (IP packets), 802.11 frames include additional fields to handle wireless-specific concerns:
| Field | Wired Ethernet | 802.11 Wireless |
|---|---|---|
| Source/destination MAC | Yes | Yes (plus additional address fields for AP relaying) |
| Frame Check Sequence (error detection) | Yes | Yes |
| Duration/NAV field | No | Yes — reserves the channel to reduce collisions |
| Sequence control | No | Yes — handles fragmentation/reassembly and duplicate detection |
| Retry flag | No | Yes — indicates a retransmitted frame |
This is why a wireless network interface card and driver must translate between 802.11 frames over the air and standard Ethernet frames once data reaches the wired portion of the network — a process largely invisible to applications and users.
Why Wi-Fi Speed Varies So Much in Practice
Several wireless-specific factors affect real-world throughput far more dramatically than on wired Ethernet:
| Factor | Effect |
|---|---|
| Distance from AP | Signal strength drops, forcing lower, more robust (slower) modulation rates |
| Interference (other networks, microwaves, Bluetooth) | Increases retransmissions, reducing effective throughput |
| Number of connected clients | Shared medium — more clients means more contention for airtime |
| Physical obstructions (walls, floors) | Attenuates signal strength |
| Channel width (20/40/80/160 MHz) | Wider channels allow higher speeds but are more prone to interference and overlap with neighbors |
Checking Wireless Signal and Rate on Linux
iw dev wlan0 link
Example output:
Connected to aa:bb:cc:dd:ee:ff (on wlan0)
SSID: HomeNetwork
freq: 5180
signal: -52 dBm
tx bitrate: 433.3 MBit/s
signal: -52 dBm— closer to 0 is stronger (typical usable range: -30 dBm excellent, -70 dBm marginal, -80 dBm and below very weak)tx bitrate— the currently negotiated transmission rate, which adapts dynamically based on signal quality
Real-World Example: Diagnosing Poor Wi-Fi Performance
iw dev wlan0 scan | grep -E "SSID|signal|freq"
This reveals all nearby networks, their channels, and signal strengths — useful for identifying channel congestion (many networks crowded onto the same channel) as a cause of poor performance, distinct from simple signal weakness.
iw dev wlan0 station dump
On an access point, this shows connected clients and their negotiated rates, useful for identifying a specific struggling client versus a broader AP-wide problem.
Comparison Table: Wired vs. Wireless Ethernet
| Aspect | Wired Ethernet | Wireless Ethernet (Wi-Fi) |
|---|---|---|
| Medium access | Full-duplex switching, effectively collision-free | CSMA/CA, half-duplex, shared medium |
| Typical speed consistency | Very consistent | Highly variable (distance, interference, contention) |
| Physical security | Requires cable access | Signal extends beyond physical walls |
| Encryption | Not inherent to Ethernet itself | Built into the standard (WPA2/WPA3) |
| Setup complexity | Simple (plug in cable) | More complex (channels, authentication, interference management) |
Best Practices
- Use 5 GHz (or 6 GHz where supported) for higher-throughput needs, reserving 2.4 GHz for range-sensitive or legacy devices.
- Choose non-overlapping channels (1, 6, 11 on 2.4 GHz) when manually configuring access points to minimize interference with neighboring networks.
- Monitor signal strength and connected client counts on busy networks — poor performance is often a contention problem, not a raw bandwidth problem.
- Use WPA2/WPA3 encryption as standard practice, understanding it’s built directly into the association process itself.
- Consider wired connections for latency-sensitive or high-throughput fixed devices (like servers, gaming consoles) even when Wi-Fi is available, since wired Ethernet remains far more consistent.
Troubleshooting
Problem: Wi-Fi speed is much lower than the advertised standard’s maximum
Check signal strength and channel congestion:
iw dev wlan0 link
iw dev wlan0 scan | grep -c SSID
A large number of nearby SSIDs on the same channel indicates congestion, not a hardware fault.
Problem: Frequent disconnects/reconnects
Could indicate marginal signal strength, interference, or power-saving mode issues:
iw dev wlan0 get power_save
Problem: One client on the AP is fine, others are slow
Check per-client statistics on the AP:
iw dev wlan0 station dump
A struggling client, rather than the AP itself, is often the actual bottleneck.
Problem: Can’t determine why throughput varies so much over time
Correlate performance drops with time-of-day/usage patterns — often caused by neighboring networks’ activity on the same channel, rather than anything wrong with your own equipment.
Conclusion
Wireless Ethernet (Wi-Fi/802.11) achieves compatibility with traditional wired Ethernet at the data layer while solving a fundamentally different physical problem: how to reliably share a single, open-air medium among many devices without wires to isolate them. Through mechanisms like CSMA/CA collision avoidance, beacon-based network discovery, and adaptive modulation rates based on signal quality, Wi-Fi delivers remarkably reliable connectivity — but understanding these underlying mechanisms is essential for diagnosing the performance variability and interference issues that wired networks simply don’t experience in the same way.
