The difference between the local network transfer speed and the displayed link speed is significant.

This thread has been locked for further replies. You can start a new thread to share your ideas or ask questions.

The difference between the local network transfer speed and the displayed link speed is significant.

This thread has been locked for further replies. You can start a new thread to share your ideas or ask questions.
The difference between the local network transfer speed and the displayed link speed is significant.
The difference between the local network transfer speed and the displayed link speed is significant.
2023-09-02 12:16:30 - last edited 2023-09-02 12:17:42
Model: Archer AX55 Pro  
Hardware Version: V1
Firmware Version: 1.2.2 Build 20230803 rel.35127(5553)

Hi all!

 

Currently, the router's WAN is set to a 1G port, and there are three devices within the environment. Specifically, one PC is wired to a 2.5G port, one NAS is wired to a 1G port, and a laptop is connected to WIFI(ax/5GHz).

In the present scenario, when I attempt to conduct iperf speed test on the local network from the PC (2.5G port) to the NAS (1G port), It can achieve near wire speed (approximately 95x Mb). However, when using the laptop (WIFI) to perform iperf speed test to the NAS (1G port), It can only achieve around 5xx Mb, even though the displayed link speed is 2Gb. There's no channel interference within the environment, and the laptop's NIC is Intel AX201. I am aware of the overhead from the physical layer to the application layer, but this difference is just too significant. Is the device performance normal?

 

Wireless test

Wired test

 

Connection info

  0      
  0      
#1
Options
2 Reply
Re:The difference between the local network transfer speed and the displayed link speed is significant.
2023-09-02 16:47:01 - last edited 2023-09-02 16:48:43

  @Soses 

 

Hi,

 

For comparison. 

 

I have an Archer AX50 that is based on Intel chips, whereas your AX55 Pro is built on chips from Qualcomm. So, the two routers are completely different.
The Wi-Fi card I am using is an Intel AX200.

 

If I run iperf3 with the same options as you did, then I also get the same low test result like you.
I am not sure if this is caused by the Intel wireless card or due to how iperf sends TCP data packets.

 

You should be able to get better results when transferring several TCP data streams in parallel. For example, add the option "-P 10" to your iperf3 command and the results should go up to > 800 Mbits/sec.

 

If we look at a more practical task. Copying a file from the NAS to the PC (i.e. the "download" direction) I get around 940 Mbps, while in the "upload" direction (copy file from PC to NAS) the throughput fluctuates quite a lot between 600 Mbps and 940 Mbps.
However, the maximum of 940 Mbps that I get is likely due to the Archer AX50 only providing Gigabit Ethernet ports.

 

Anyway, if your results are about the same as mine or higher, then I would say there is nothing wrong with your AX55 Pro.

 

While Wi-Fi cards based on the tiny little Intel AX200, AX201, AX210 chips are the fastest that normal people like us can currently buy, they might not be truly as powerful as their specifications suggest.

 

  0  
  0  
#2
Options
Re:The difference between the local network transfer speed and the displayed link speed is significant.
2023-09-02 18:03:18

  @woozle 

 

Hi, thanks for the response.

 

When using parallel transfers, I indeed achieve speeds similar to yours, reaching up to 9xx Mb.

However, when I add the -R option (reverse mode), the speed remains just below 600 Mb. Yet, under wired connections, I can still achieve up to 9xx Mb in this mode.

 

In the context of file transfers, I get results opposite to yours.

When I tried downloading a file from the NAS to the laptop, the transfer speed was still only around 5xx Mb (which actually aligns with the previous test, where in reverse mode, the server sends packet traffic to the client). But when uploading, I can reach around 8-9xx Mb.

 

I think this download speed a bit puzzling...

  0  
  0  
#3
Options