WiFi Jitter Outlier Versus Ethernet Performance | Generated by AI
Here’s what the data tells us:
ANALYSIS: Network Latency to 192.168.1.36
==========================================
Your Air is on Wi-Fi (en0), IP 192.168.1.52
Both .1 (router) and .36 (workstation) appear as Wi-Fi neighbors in ARP
(both on en0 — .36 is actually wired, but router bridges the LAN)
PATH: Air [WiFi] -> Router [Ethernet] -> Switch -> Workstation (.36)
MEASUREMENTS (30 pings each):
min avg max stddev
Router (WiFi only): 7.4ms 22.9ms 96.0ms 26.8ms
Workstation (full path): 4.4ms 12.4ms 69.5ms 15.8ms
KEY INSIGHT: Your 92.3ms was an OUTLIER, not the typical case.
The workstation actually has LOWER avg latency than the router!
(12.4ms vs 22.9ms — WiFi contention varies per-packet)
LATENCY BREAKDOWN (estimated):
WiFi hop (Air -> Router): ~5-15ms typical, spikes to 50-90ms
Router processing: ~0.1ms
Ethernet (Router -> Switch): ~0.01ms (speed of light in copper)
Switch -> Workstation: ~0.01ms
WHY THE SPIKES (92.3ms, 74.9ms, etc.):
- WiFi channel contention (neighbors on same channel)
- WiFi power-save mode wake-up latency
- 802.11 retransmissions on interference
- Background scan for better APs
- All of these add 10-100ms jitter to WiFi
VERDICT: The 92.3ms spike is pure WiFi jitter.
The Ethernet/switch leg adds <1ms of deterministic latency.
Your table makes sense — the extra path (router->cable->switch)
is negligible. WiFi is the bottleneck, not the topology.
Compare: if .36 had WiFi too, you'd see 2x WiFi jitter (40-200ms).
Being wired, only one WiFi hop contributes.
The 92.3ms you saw was a WiFi jitter spike, not the typical case. With 30 pings, the average is only 12.4ms — the Ethernet/switch path adds essentially zero latency. The --delay flag captures single-sample latency, so it naturally picks up these WiFi spikes. For more stable measurements, you’d want something like ping -c 10 averaged. But for a quick network scan, it’s useful — if you see consistently high latency to a device, that tells you something real.