Help Proper Config GT-AX11000 2.5ge + 2.5ge switch + 2.5g modem for over 1gbps...

[thecolin85]

Hi all, newbie here, please be patient with me. I have a "weekend-warrior" level knowledge of tech stuff, but I'm no networking expert. Otherwise I probably wouldn't be asking these questions…
I cannot get my current configuration to provide my 1200mbps+ speeds across multiple devices, even after doing all this research months back and thinking, "Oh if I just buy a multi-port 2.5g switch, I won't have all my devices throttled below 1gbps. And I had thought I had it figured out at one point. I swore I had it once where if I plugged my desktop PC in hardwired, I would get those 1300mbps+ speedtest results, but on WiFi it would NEVER break 1000mbps. Always 900-something mbps. I tried tweaking the router settings, maybe sure all drivers and firmware all up to date etc. No dice.

My hardware:

• Modem: (Motorola) ARRIS Surfboard S33 DOCSIS 3.1 Multi-Gigabit Cable Modem with 2.5 Gbps Ethernet Port
  
• 
  
• Router: ASUS ROG Rapture (GT-AX11000)
  
• 
  
• Switch: TP-Link TL-SG105-M2 5-Port 2.5G Desktop Switch
  

Also running the router on the latest firmware of Asuswrt-Merlin. Desktop PC is an x570 Crosshair Hero VIII Wi-Fi (FWIW).

I have tried so many different configurations (using at least CAT6A, CAT7 or CAT8 cables) of what to plug into what first. Currently I have the 2.5g port out of the S33 modem and into the 2.5g port on the AX11000 router. Then out of the AX11000 router's blue (WAN?) port and into one of the four 2.5g ports on the switch. Then out of the switch, I have a 2.5g port and cable running to my desktop PC's 2.5g onboard ethernet port.

I verified I can get 1300 and nearly 1400 mbps speeds when I run the 2.5ge port on the S33 modem straight to my desktop PC. I've also verified I can break over 1gbps Wi-Fi speeds on various PC devices (desktop, laptop) and mobile phones when plugged into the 2.5g port on the router, but then anything hardwired is capped at 900-something MBPS.

I don't know if I should run the modem 2.5ge straight to the router first, then to the switch? Or to the switch, then back to the router… Or am I screwed or limited by the hardware somehow? Should I run a cable from the S33 modem's 2.5ge port to the router's 2.5ge port and then run another cable from the S33 modem's 1ge port to the blue "WAN" port on the AX11000 router?

I could care less if all Wi-Fi devices were capped below 1000 mbps / 1gbps as long as I could have at least my own desktop PC device able to get over 1000mbps when hardwired especially.

[TonyDTiger]

@BertoAndrade - That makes sense for the iPhone 14 Pro Max.  The iPhone 15 Pro Max is the first Apple device to be able to use the Wifi 6E, 6Ghz band.. to which you can see it clear from my screen shot that I connect to my 6Ghz SSID - I am able to speedtest beyond the 1Gbps threshold wireless.

@jzchen - I did also try just now switching to USB Legacy mode and it didn't seem to make a significance difference compared to my earlier postings when using that device as a WAN source either.  Appreciate the tip tho! 

At the moment, my best internet speeds received for the router and for all downstream clients are when I am using the 2.5Gbps WAN port but again open to any suggestions to see if there is something on the router settings side that could enable speeds higher than 750Mbps/1250Mbps with a carrier provided 2Gbps/2Gbps

[jzchen]

I'm not sure if @BertoAndrade 's point is clear to you, but he is getting almost double download speed on computer clients, than what is showing up on his router (built-in) speed test.

In that light, I am able to pull faster download speed on my computer via 10 GbE backhaul than from my router (built-in) speed test.  (See screenshots).  We are hoping you can experience the same effect, and proposing possible ways to test/check based on your setup...

EDIT:  I'm not sure if my point was clear, USB 3.0 does produce some WiFi noise.  This noise coincides with data stream flowing through.  Unfortunately, there is a peak waveform just starting from 5.0 GHz which is my concern is limiting your WiFi speed, hence trying on 2 ethernet/wired clients...

[TonyDTiger]

Thanks for clarifying @jzchen, while I do get that but even with your speed tests provided, it doesn't appear that you're using the same server to test.  I've noticed that the ASUS router tends to pick a certain speediest server (in your screenshot it looks like it's hosted by Frontier) while the speediest.net ookla service tends to select another (Giggle Fiber).  It would be too to select the same server to ensure it is Apples to Apples.  I have seen in some cases when selecting certain servers I get grossly false readings.  For example, if I select the COX server it may say 40mbps download with 1500mbps upload.  Super weird right?  In some of my testing I tested 3x across different servers in these drop down lists  to have a higher confidence that the reporting was accurate too.

I think that along with turning off AIProtect + QOS on the router while conducting these is important, then of course re-enable if preferred.  In my test cases above and even outside of reported here, I never saw a time where a downstream client received a higher speedtest than the router itself, although I could potentially see that possible for other users if they have QOS turned on and they had prioritized that particular device.  IMO QOS works really well when you have many bandwidth hungry devices with limited service bandwidth but I don't think it's needed in my case since I am not doing anything at a crazy commercial level with the bandwidth I have available lol.  So I have QOS off entirely because of this..  If it's worth mentioning, when I do test - I do turn off AiProtect, but outside of that I do turn it back on.  That being said though, AiProtect has a marginal impart to my Speed Tests, but not enough to really to be impactful to my concerns mentioned here. 

I hope that adds context to everything but if I am still misunderstanding something - please by all means feel welcome to virtually smack me on the head with a follow up reply while kindly ELI5 haha.  Thank you all kindly~

[jzchen]

I didn't catch that the server was changed (due to Auto server choice built into speedtest.net), so I'm afraid that I do have to concede to the apples to apples remark, because of that.  That said, the network conditions are never going to be apples to apples, given two different moments in time.  The auto choice is supposed to select the best server at any particular moment.

I did a couple of speedtests in this manner.  I logged into ASUSwrt on my phone, 192.168.50.1, and turned on my computer, (then manually selected the same server on both).  I simply clicked start on both devices at (about) the same time.  (Those are the results towards the top on the router recorded speedtests).  The speed tests don’t stop dowload at the same time, the one that ends download first tended to have poorer results as bandwidth was released to the other, but during the time both were concurrently running I could see total throughput divided to approximate sum to my total maximum, as expected.

Theory is theory, and practice is practice.  Please note, I never get into the 11xx on the router built in speedtest.

If, you would so kindly give @BertoAndrade the benefit of the doubt, yours should not be halved during the simultanous test...

EDIT-  AiProtection is always on, when QoS accidentally got turned on, I could never replicate the high speeds…

[TonyDTiger]

Ok.  So to properly follow his guidance, I will need to purchase another USB to LAN device as the only devices that have multi-gig Ethernet capability include:

1 AXE11000 router (only device capable of broadcasting 6GHz)

2 XT8 satellite nodes (where backhaul is being used)

3 goCoax 2.5 mocha adapters (for above backhaul)

1 ASUS 2.5G USB 3.2

To be clear, he had mentioned it was unfavorable to determine wireless reporting results, even though I had demonstrated:

• AXE11000 connected to modem via 2.5G, Router speedtest ~2Gbps, iPhone 15 Pro Max connected to 6GHz SSID was able to receive 1.5Gbps
• AXE11000 connected to modem via USB,Router speedtest ~700/1500Mbps, iPhone 15 Pro Max connected to 6GHz SSID was able to receive 300/600Mbps..

This will take a few days before this comes in the mail and I’m going out of town this morning but will report back when I am able to test under these conditions.

[BertoAndrade]

No need to buy another USB. Do you have any other computer at home? If so, what @jzchen suggested is to just go ahead and run speedtests at both at the same time or at a computer and at your phone at the same time, ignoring the servers they’re testing from.

From what I understood, he was able to get higher speeds on the clients as opposed to the router speed tests, same scenario as myself.

[TonyDTiger]

Ok - I see the half but it is worse than the whole if I am interpreting this correctly?  I made two videos of this test case scenario in the attached link.  One with USB connection another with 2.5Gbps LAN.  I hope you are able to access the link: https://photos.app.goo.gl/Vvv2GGYZzGhHCRaR8 and please keep in mind my MacBook and iPhone both are 6Ghz enabled and connected while doing this...

[BertoAndrade]

There is indeed something different in your scenario. My speed tests combined always gave me higher results than the router speed tests alone, reaching the full connection speed. Have you tried a factory reset on the router? I know how much of a headache it can be to set everything back up, but sometimes it’s all it takes to fix a few issues.

[TonyDTiger]

Do you have QOS turned on?

[BertoAndrade]

Yes. Everything is turned on in my router: QOS, Trend Micro Security and the AIMesh. It doesn’t make any difference in the speed tests if I turn them off.