A point of curiosity about Wifi 802.11 ac overhead. I have a wired Ethernet Internet link which tests at around 500 megs download. Using a Wifi 802.11 connection, the download speed tests around 285 megs download. Am I correct in assuming the difference in download speeds is typical of WiFi overhead?
WiFi overhead 802.11 ac
- Thread starter lillard
- Start date
Solution
ShadyHamster
Distinguished
- Jan 3, 2012
- 1,925
- 3
- 20,960
If there are other devices connected to the Wifi and you aren't sitting on top of the router for 100% signal then your Wifi speeds will be slower then advertised.
The overhead on Wifi shouldn't be any more or less then wired networking.
The overhead on Wifi shouldn't be any more or less then wired networking.
BFG-9000
Splendid
- Sep 17, 2016
- 3,424
- 440
- 27,090
The overhead of gigabit ethernet is 6% so efficiency is 94% with standard 1518 byte frames.
WiFi is much more complicated than it seems because its advertised peak modulation schemes such as 3/4 256-QAM barely work under the best of conditions so are likely to fall back to slower ones in normal use >10' away. The signal strength required goes up with both wider bands (e.g. 160MHz wide) and bonding more antennas (2x2, 3x3) so you are even less likely to connect at full speed. Any interference at all will result in lost packets and retransmits. Furthermore, while AC takes less airtime to send the data, preambles, ACK packets, RTS, CTS and the CSMA protocol take just as long as before for backward compatibility so while N had ~85% theoretical efficiency, with AC it's down to 65%. In general you are doing well to reach even 1/3 of the claimed link speed at any reasonable distance from the AP.
The sad thing is WiFi manufacturers have thrown in everything they could to get large bandwidth numbers to advertise, even at the expense of latency which is the more annoying problem. FQ_CoDel fixed bufferbloat latency issues across the wired internet fairly well, and we can only hope the Airtime-fair-FQ project from the same people will help do the same with WiFi.
WiFi is much more complicated than it seems because its advertised peak modulation schemes such as 3/4 256-QAM barely work under the best of conditions so are likely to fall back to slower ones in normal use >10' away. The signal strength required goes up with both wider bands (e.g. 160MHz wide) and bonding more antennas (2x2, 3x3) so you are even less likely to connect at full speed. Any interference at all will result in lost packets and retransmits. Furthermore, while AC takes less airtime to send the data, preambles, ACK packets, RTS, CTS and the CSMA protocol take just as long as before for backward compatibility so while N had ~85% theoretical efficiency, with AC it's down to 65%. In general you are doing well to reach even 1/3 of the claimed link speed at any reasonable distance from the AP.
The sad thing is WiFi manufacturers have thrown in everything they could to get large bandwidth numbers to advertise, even at the expense of latency which is the more annoying problem. FQ_CoDel fixed bufferbloat latency issues across the wired internet fairly well, and we can only hope the Airtime-fair-FQ project from the same people will help do the same with WiFi.
Facebook
Twitter
Instagram