The "powerline is low latency" myth

[skywake]

A more detailed graph with quite a bit more data, I'll update the first post with it:

[Tim Drake]

Surely something like a Powerline is highly dependant on your internet, power lines AND  what adapter that you're using..

[skywake]

Surely something like a Powerline is highly dependant on your internet, power lines AND  what adapter that you're using..

Internet: No, not for your internal network. I'm pinging my router not google or my ISP. Also with latency every bit of latency in the chain adds up.

Wiring: Yes. And I'm sure some people's wiring is so bad it's completely unusable. But nobody is posting better numbers

Adapter: For bandwidth sure, better spec adapters do quite a bit better. For latency? ... well, nobody has posted better numbers

[arronleeds]

here guys something to  see

 

powerline

 

 

 

 

5Ghz @ 300Mbps

 

 

 

i ignore speedtest pings, i dont get 13ms on uk servers to bf4 with powerline but i do get 11ms to uk servers on wireless.

[System Error Message]

This thread may be old but i thought i would chime in some facts for those who are comparing wifi and powerline.

First lets look at conditions. Powerline is affected by how good your wiring is and what you run. Wifi is affected by traffic and obstacles.

Both powerline and wifi are shared mediums, even if you have the AV2000 with its MU-MIMO the copper is still shared however wifi is a shared space so anything running on the same channels affects other devices on the same channels too. Its easier to keep powerline in check than it is with wifi.

 

With the best of the powerline adapters i still get 3-5ms latency, however that latency is consistent even during heavy loads. With wifi the idle latency can creep below 1ms using broadcom's MU-MIMO AC3100 on 5Ghz AC wifi with Intel AC wifi NIC. When the medium is used more the latency increases much. Even with ethernet filling up a 1Gb/s link the latency does increase significantly when the link is fully used. With wifi you have to worry about other wifi traffic on the same channels that dont belong to you.

 

Ethernet is the most consistent followed by powerline than wifi. A more scientific test needs to be done since the first post was done on idle. For a proper test even speed test isnt valid because the ping isnt done while the test isnt conducted. The best way to actually measure the performance is to fully use it and than ping. For example transferring files from 1PC to another and pinging it at the same time. In my case i uploaded files filling up the full 1Gb/s NIC with a nice straight line and pinging it, ping times varied between <1ms to a few ms. I tried the same with wifi, filling up the link till it get to 90% link utilisation and packets were dropping.