Available Now: Buy the new iPhone 14, iPhone 14 Pro and iPhone 14 Pro Max from AT&T!
Get superfast AT&T Fiber internet
dannyloski's profile

New Member

 • 

12 Messages

Mon, Jun 27, 2022 1:42 PM

AT&T Internet 1000 Fiber Plan, but only getting 100-200Mbps via Ethernet

I just switched over yesterday from Spectrum to AT&T Fiber as it was just made available in my neighborhood (been waiting patiently for this lol). I went with the AT&T Internet 1000 plan, which offers 1Gbps speeds.  However, while my PC is connected via Ethernet I'm only getting 100-200Mbps.  Doing a speed test at https://www.att.com/support/speedtest/ gives me the below results:

So my gateway seems fine as it's getting the 1Gbps, but for some reason my PC is not getting anywhere close to that. Here's more info on the setup ...
 
Computer:
Lenovo Thinkpad X1 - Gen 1 ( specs)
Windows 10 64-bit
Intel Core i7
8GB RAM
Lenovo Thinkpad USB 3.0 Pro Docking Station ( specs)
 
Connection:
Ethernet Cable (Cat6a) connected from Port 4 of gateway to the Lenovo Thinkpad USB 3.0 Pro Docking Station.  This docking station features a Gigabit Ethernet Port and is connected via USB 3.0 to the PC.  The Network Adapter properties page shows the link speed of 1000/1000 Mbps.
 
AT&T Fiber Gateway:
Model BGW320-505
Firmware 3.17.5
 
Troubleshooting Done:
  • I tried a few different Cat5e, Cat6, and Cat6a Ethernet cables, but problem persists.
  • Set Ethernet Adapter's settings to 1.0 Gbps Full Duplex as well as Gateway's Port 4 settings to 1G full duplex, but problem persists.
  • Tried different ports on the back of the Gateway, but problem persists.
  • Removed docking station and connected PC via USB 3.0 to Gigabit Ethernet Adapter dongle (specs), but problem persists.

 

Any helpful tips or insights to try and get better speeds on my PC?  Appreciate any and all help.  Thanks in advanced!

JefferMC

ACE - Expert

 • 

29.8K Messages

3分前

1) Do you have AT&T Internet Security enabled in Smart Home Manager?  Have you tried turning that off?

2) There have been reports of reduced performance when using the Ethernet port on the USB3 docking station, vs. a native Ethernet port.  While using a different USB3 dongle does bypass any issues with the docking station, it doesn't eliminate the USB3 adapter overhead.  While USB3 is theoretically capable of 5 Gbps for sustained transfers, that doesn't mean that Gigabit Ethernet adapters will automatically achieve that.  You might try local transfers within your network to see if these adapters are capable of full Gigabit performance.

3) Did you disable the Wi-Fi connection while testing wired, or is it possible it's using both?  That jitter seems rather high, and is indicative of Wi-Fi.

4) Does Lenovo have any "Network Accelerator / QoS" software in their network driver package?  Do you have that enabled?

(edited)

New Member

 • 

12 Messages

3分前

@JefferMC thanks for the prompt reply.  Hopefully, I can answer the questions you asked.

1. Yes, I have AT&T Internet Security enabled in Smart Home Manager.  I just turned it off and the speed actually worsen.  See picture for new Speedtest results.

 

2. Yes, I'm aware that speed can be impacted when using Ethernet over the USB 3.0 Docking/Dongle versus a native Ethernet port.  However, I would assume it shouldn't be by this much.  As you said, USB 3.0 should be able to handle up to 5Gbps transfer speeds.  Unfortunately, my PC does not have a native Ethernet port to test on.

 

3. Yes, Wi-Fi connection was disabled during the Ethernet connection speed tests.

 

4. I'm unsure, can you provide more details where I can go to validate this?

(edited)

JefferMC

ACE - Expert

 • 

29.8K Messages

3分前

1) That's rather unexpected.  Your jitter got better, though.  

2) I was thinking that if you had a NAS or another PC that had a Gigabit adapter you could try doing a direct transfer via your USB3-based adapters to see what you got.

3) Cool. 

4) I'm not up on Lenovo's bundles and whether they have Network accelerators included and what it might be called. Perhaps someone else can pop in if they know.  You could look through the list in Add/Remove programs for things that you have no idea what they might be, especially if they have the work "Network" in them.

New Member

 • 

12 Messages

3分前

@JefferMC appreciate you chiming in.  I'll wait to see if anyone else has any tips.  Thanks again!

JefferMC

ACE - Expert

 • 

29.8K Messages

3分前

If you haven't so far, you might want to try doing a factory reset on that Gateway.

New Member

 • 

12 Messages

3分前

@JefferMC I hadn't done a factory reset, seeing as the installation was literally done yesterday.  However, I figured why not and just completed the factory reset on the Gateway ... Here's the speed test results:

tonydi

ACE - Guru

 • 

7.3K Messages

3分前

Unfortunately, I don't have any additional suggestions to offer, Danny. 

But I wanted to say that if we had a way to pin a post in these forums and make people needing technical help look at them before creating their own, yours would be the best example I could choose.

Absolutely stellar job of providing everything necessary and laying it all out in a clear manner.

👍👍

JefferMC

ACE - Expert

 • 

29.8K Messages

3分前

Absolutely stellar job of providing everything necessary and laying it all out in a clear manner.

Agreed. 

I hadn't done a factory reset, seeing as the installation was literally done yesterday.

You wouldn't think a newly installed gateway would need a factory reset, but in many cases you would be wrong.  I'm not sure what the explanation is (I have several candidate reasons, more than one may be true), but it a factory reset fixes a host of ills.

Well, your upload has gone up.  That's a start.  17 ms latency seems a trifle high for fiber.  Would you mind sharing what town or at least county you're in?  Also, a traceroute to that speed test site would be lovely.  If you'd like, I'm going to include a bit of PowerShell you could run to see where your top 10 test hosts are:

    $uri = 'http://beta.speedtest.net/speedtest-config.php'
    try {
        [xml]$config = Invoke-WebRequest -Uri $uri -UseBasicParsing -ErrorAction Stop
    }
    catch {
        Write-Error "Could not download configuration from $uri"
        Break
    }

    Write-Output "Testing from $($config.settings.client.isp)"

    Write-Output 'Retrieving server list...'

    $uri = 'http://www.speedtest.net/speedtest-servers.php'
    try {
        [xml]$hosts = Invoke-WebRequest -Uri $uri -UseBasicParsing -ErrorAction Stop
    }
    catch {
        Write-Error "Could not download server list from $uri"
        Break
    }

 $hosts.settings.servers.server | select sponsor, name, host

Why do that?  So you can pick out your host and do a traceroute to it.  However, it occurs to me here that when doing a speed test on an Fiber connection, it typically needs to be multithreaded in order to get fiber speeds.  That typically means more than one host is used.  And Speedtest.net only discloses one.

Another suggestion would be to run the SpeedTest Windows App to see if it gives different results.

tonydi

ACE - Guru

 • 

7.3K Messages

3分前

🤔  Not sure what Ookla is doing there.  The PS script delivers a list of ten sites, only five of which show up in the apps top ten choices.  Some of the sites in the script list are hundreds of miles away yet it doesn't show the AT&T server that is 3 miles away.

Yet if you just go to the URL in the script, that result doesn't even match the one the script delivers (but does have the local AT&T server).

New Member

 • 

12 Messages

3分前

Thanks guys! I appreciate the kind words.  I figured I'd provide as much information as possible from the get-go, so that you all have a full view of the scenario and not be as blind when trying to provide feedback and help.

 

I did see an improvement on the Upload after the Gateway was factory reset, like you said, that's an improvement! Unfortunately, not so much on the download.  As for where I'm located, it's in Orlando, Florida.

 

Here's the output for the PowerShell:

 

 

Here's the TRACERT result to the Century Link host which is what was selected by default when I did the Speedtest through https://www.att.com/support/speedtest/

 

 
Now here's a very interesting result. I downloaded the Speedtest App from the Microsoft Store and ran it.  Take a look at the results:
 
 
Then I went via Speedtest.net and performed the test to the same server and look at those results.
 
JefferMC

ACE - Expert

 • 

29.8K Messages

3分前

Not sure what Ookla is doing there. 

Toni, yeah, that's code I boiled out from a PowerShell project that mirrored the similar code in a JavaScript project I had just looked at.  I notice my returns are odd today, but this is the first time I've noticed it not being spot on what the web apps does.  Perhaps they've been playing.  These projects usually compute distance to the various servers and pick a handful of close ones to actually ping.  Not that it will make much difference because...

Here's the TRACERT result to the Century Link host 

Well, the first thing I notice about your traceroute is that you get routed from Orlando to Atlanta to get back to a server in Orlando, which is just craziness.  Though that could be the way peering works in Florida.  I would think there would be peering in Miami or somewhere down there closer to Orlando.  There are a few major data centers around Orlando.   Click on the server selection and chose something in Miami (maybe it'll peer there) or near Atlanta (to keep you from having to make two round trips).

UPDATE: On your last ones I see your PING dropped to 5, telling me that it's local peering at least.  Still, things don't seem sane.

(edited)

New Member

 • 

12 Messages

3分前

Ok, so this is weird ... I went with Frontier in Miami and ran the TRACERT for it.  Here's the result:

 

 
I then selected Frontier as the server via the Speedtest App I downloaded from the Microsoft Store and got these amazing speed results (that's more like it!).
 
 
However, if I select the same server via the browser version of Speedtest.net I get these results:
 
 
Why?!  That's a very huge difference in speeds and I'm unsure what is going on.  Which of the results is the actual result of speed I'm actually getting?

(edited)

JefferMC

ACE - Expert

 • 

29.8K Messages

3分前

When running speed tests, the best number you get is likely the most accurate measure of your connection speed because your connection speed is only one small part of the story.  All speed tests are done by downloading a known amount of bits and measuring how long it takes them to get there.  There are many things that impact the amount of time it takes to get from point to point, including:

  • The speedtest site's server busyness
  • The speedtest site's network
  • the peering between that network and your ISP's network
  • your ISP's network
  • your actual connection
  • your gateway/router
  • your network
  • your PC hardware
  • the network stack on your PC
  • the application on your PC

Any or all of these things can make it take more time and make the measurement look slower,  It's almost impossible for any of them to make it faster.

(edited)

tonydi

ACE - Guru

 • 

7.3K Messages

3分前

Not sure what's going on with this thread.  I've made two posts that just don't show at all and now the pictures in Danny's last post are just gone.

JefferMC

ACE - Expert

 • 

29.8K Messages

3分前

the pictures in Danny's last post are just gone.

There were pictures?  I wondered about that.

BTW, I peeked into the current code; there's a different API being called now to get servers.  I will have to update my PS scripts. :)

(edited)

Need help?
New to AT&T Community?
New to the AT&T Community? Start by visiting the Community How-To.
New to the AT&T Community?
Visit the Community How-To.