DQOTD: WiFi Debottlenecking?

[ERD50]

Is that even possible - 8 Mbps even momentarily from a 1000 Mbps service? I can well understand some fluctuations, but I'd be furious if I paid for 1000 Mbps.

I meant that even if you are getting 1,000 Mbps in, that if you had a really bad/weak Wi-Fi connection, that part of the connection might drop down to 8 Mbps (the source should still be able to provide 1,000 Mbps).

I suppose even very high speed ISP could drop for a moment (fraction of a second), but your streaming video buffers up 10 seconds or so, so any short drops should have no effect on your viewing experience.

-ERD50

 

[Cobra9777]

... Again, I'm going to suggest actual tests with actual numbers. With a speed test, and a ping test, is the wireless connection measurably worse than wired, and if so, does the wireless connection drop a good speed down to a marginal speed? ...

I have Frontier FiOS 50/50. I just ran 2 speed tests (speedtest.net) on my laptop, one wired to the router, the other on WiFi. The router was about 3 feet from the laptop on the WiFi test. I shut down all other connected devices. Here are the results:

Wired: ping 9 ms, download 48.60 Mbps, upload 61.41 Mbps
WiFi: ping 12 ms, download 24.70 Mbps, upload 26.90 Mbps

I then walked over to the main TV (about 60 ft from the router) and ran another speedtest on Wifi. Here is the result:

WiFi: ping 17 ms, download 13.67 Mbps, upload 15.73 Mbps

That's a huge drop from 50/50, but it *should* still be adequate for streaming HD video. But keep in mind, that's a speedtest using a nearby telco server and small binary files. Actual performance streaming 1080p video plus 5.1 audio from across the country is going to be more sketchy because the path typically involves 15-20 hops. Plus I had shutdown all other devices on the network.

Suit yourself, but I know from many years of experience with multiple TV streaming set-ups that buffering is more common on Wifi vs cat5 to the router. And it's such a simple way to eliminate an unstable variable especially if you have a router nearby. I just recently helped DS run cat5 to the master bedroom in his new house because Netflix was completely un-watchable on Wifi. That fixed his problem.

Another recent example... we got back from a trip a few months ago and I hooked up the laptop to the main TV, which is where it normally resides running Kodi. DW was complaining about buffering for several days. After some troubleshooting, I realized the laptop was still in wireless mode. Even though the ethernet cable was plugged in, I had disabled that connection while we were traveling. Re-enabled it and boom, no more buffering.

 

[cbo111]

After months of buffering issues with Brighthouse (Spectrum), we just moved our modem to a location right next to the main tv used for Netflix, Amazon Prime, etc. We then ran an ethernet cable directly from the modem to the tv. Now we can stream with zero buffering issues. Only downside was we had to pay a $39 service charge to Spectrum for them to reroute the tele signal to accomodate the new modem location.
Chuck

 

[Sunset]

If you are using wifi, it's not just the distance, but the walls in between that can affect the speed. Even worse is floors , like if you have the router in the basement and are trying to watch tv on the 2nd floor.

Router antenna's are designed to broadcast horizontally and not so much upwards.

 

[meierlde]

If you are using wifi, it's not just the distance, but the walls in between that can affect the speed. Even worse is floors , like if you have the router in the basement and are trying to watch tv on the 2nd floor.

Router antenna's are designed to broadcast horizontally and not so much upwards.

If you look at the new routers using band ac, they have 6 or 8 antennas so they can do directional beaming making signals stronger. Plus the 5 ghz band tends to have fewer neighbors on it so more bandwidth is available for your use. If going upstairs you could hang the router on a wall. (newer models include a template to drill the needed holes.

 

[ERD50]

I have Frontier FiOS 50/50. I just ran 2 speed tests (speedtest.net) on my laptop, one wired to the router, the other on WiFi. The router was about 3 feet from the laptop on the WiFi test. I shut down all other connected devices. Here are the results:

Wired: ping 9 ms, download 48.60 Mbps, upload 61.41 Mbps
WiFi: ping 12 ms, download 24.70 Mbps, upload 26.90 Mbps

I then walked over to the main TV (about 60 ft from the router) and ran another speedtest on Wifi. Here is the result:

WiFi: ping 17 ms, download 13.67 Mbps, upload 15.73 Mbps

That's a huge drop from 50/50, but it *should* still be adequate for streaming HD video. ...

Interesting numbers. Even though, as you say, the 13 Mbps should be adequate for streaming, the fact that it dropped from 48 wired could mean that it might be susceptible to interference, or even people moving around in the house, phase of the moon, sun spots, whatever. Though I'm not sure the other upstream delays you talk about would be an issue - that might cut your raw speed from 48 to something less, but I would think as long as it stays above the ~ 13 you get over wireless, it would be OK. The wireless shouldn't change as a fraction of the incoming, it should stay at that cap set by the wi-fi connection integrity. Weakest link. I think.

-ERD50

 

[sengsational]

Not surprising at all that once you go to a 1,000 Mbps source, that some deltas in Wi-Fi will show up. You've gone from the ISP being the limiting factor, to the routers being the limiting factor (weakest link in the chain approach).

But the important point (at least for streaming HD video), is did the Wi-Fi signal drop below 8 Mbps at any of those points?

-ERD50

It turns out that it wasn't Google Fiber that gave me the insight into the WiFi speed differences, it was the same thing I've been using "forever": dslreports/speedtest. The ones that say WiFi are going through an older WiFi router that's farther away. The ones that say Fiber are going right to the Google Fiber box's WiFi.