Decreased network performance after hardware upgrade

[Whattteva]

On the server, I use the OpenSpeedTest plugin.

Not familiar with this, so I can't comment.

On the laptop, I use the speedtest by Ookla.

This test isn't really accurate cause it only measures the speed of your laptop connection with the internet, not to your server, which is what matters.

For the video transfer from the server to my desktop, I look at the file transfer speed.

If this is through SMB, and with files you've used in the past, it's possible that most of the transfer has already been cached and you're getting much faster speeds.

Also worth noting, you could get different speeds on iperf depending on which one is the server and which one is the client. See what happens when you reverse the two commands.

 

[Whattteva]

I also tried an iperf between my desktop and my laptop, still the same result, 80mb/s, yet both are connected in 1GB/s, I checked.

It's important to get the terminologies correct. I think you are somehow mixing them up. Use little "b" for bits and big "B" for bytes. For reference, 1 Byte is 8 bits. You can see by using wrong terminologies, you are off by a factor of 8.
Your 1 GB/s is inaccurate and is a different quantity from 1 Gbps. You probably meant 1Gbps which translate to roughly ~900 MB/s.

 

[Proxipupuce]

Voyez ce qui se passe lorsque vous inversez les deux commandes.

always the same result

You probably meant 1Gbps which translate to roughly ~900 MB/s.

yes, I meant that. I know the difference, it was a mistake on my part. Sorry!

 

[Proxipupuce]

You probably meant 1Gbps which translate to roughly ~900 MB/s.

1Gbps is 900 Mb/s, not 900MB/s, right?

Apart from all that, the problem would come from the Realtek controller?

 

[Whattteva]

1Gbps is 900 Mb/s, not 900MB/s, right?

Oh yes, my mistake that was actually a complete mistake, good catch:
Here's the correct conversions:
1Gbps = 1'000'000'000 = 1'000'000'000 / 8 / 1024 = 119 MiB/s
Note that MiB (MebiByte) is what typically OS's use to measure speed.
This is slightly different from MB = 10E6 * 8 = 8'000'000 vs MiB = 1024^3 * 8 = 8'589'934'592

Apart from all that, the problem would come from the Realtek controller?

I wouldn't be so fast on that. It could be, but if you have a third computer apart from your laptop and your server, I would perform iperf test between that third computer and your laptop and also between that third computer and your server.
The results between all those tests would be more give you a better idea on where really is the bottleneck since presumably they all have different NIC's.

 

[Proxipupuce]

Oh yes, my mistake that was actually a complete mistake, good catch:
Here's the correct conversions:
1Gbps = 1'000'000'000 = 1'000'000'000 / 8 / 1024 = 119 MiB/s
Note that MiB (MebiByte) is what typically OS's use to measure speed.
This is slightly different from MB = 10E6 * 8 = 8'000'000 vs MiB = 1024^3 * 8 = 8'589'934'592


I wouldn't be so fast on that. It could be, but if you have a third computer apart from your laptop and your server, I would perform iperf test between that third computer and your laptop and also between that third computer and your server.
The results between all those tests would be more give you a better idea on where really is the bottleneck since presumably they all have different NIC's.

I did what you told me, and the results are amazing.
I took another laptop, on Wi-Fi, and :

[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 28.8 MBytes 241 Mbits/sec 0 208 KBytes
[ 5] 1.00-2.00 sec 33.2 MBytes 279 Mbits/sec 0 208 KBytes
[ 5] 2.00-3.00 sec 33.6 MBytes 282 Mbits/sec 0 208 KBytes
[ 5] 3.00-4.00 sec 36.2 MBytes 303 Mbits/sec 0 208 KBytes
[ 5] 4.00-5.00 sec 32.2 MBytes 270 Mbits/sec 0 208 KBytes
[ 5] 5.00-6.00 sec 34.2 MBytes 287 Mbits/sec 0 208 KBytes
[ 5] 6.00-7.00 sec 36.4 MBytes 305 Mbits/sec 0 208 KBytes
[ 5] 7.00-8.00 sec 36.5 MBytes 306 Mbits/sec 0 208 KBytes
[ 5] 8.00-9.00 sec 32.8 MBytes 275 Mbits/sec 0 208 KBytes
[ 5] 9.00-10.00 sec 34.7 MBytes 291 Mbits/sec 0 208 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -

On this pc, too, I reach 50MB/s in SMB sharing instead of 20!
This is still low for 1Gbps, but it is better.

 

[Proxipupuce]

On the other hand, between this laptop and my desktop, it is always 70mbps...
While when I download a game or anything on my desktop, it's 400mbps

 

[Whattteva]

I did what you told me, and the results are amazing.
I took another laptop, on Wi-Fi, and :

On this pc, too, I reach 50MB/s in SMB sharing instead of 20!
This is still low for 1Gbps, but it is better.

Those results are actually pretty typical for WiFi. You're not gonna be getting Gigabit speeds on WiFi for sure, but anything between 250-500 Mbps on WiFI is basically almost as good as it gets barring no RFI from other sources and being literally right next to the antenna.

 

[Whattteva]

On the other hand, between this laptop and my desktop, it is always 70mbps...
While when I download a game or anything on my desktop, it's 400mbps

Can I get a clarification, which computer is which? The terms desktop and PC are kind of confusing me. You have a server, a desktop, and two laptops?

Also, which ones are giving you poor iperf results?

 

[Proxipupuce]

Can I get a clarification, which computer is which? The terms desktop and PC are kind of confusing me. You have a server, a desktop, and two laptops?

I tend to speak fast thinking I'm understandable, but that's often not the case, sorry!
Let me clarify:

- My server
- My Desktop (fixed computer in my room)
- Two laptops

 

[Whattteva]

I tend to speak fast thinking I'm understandable, but that's often not the case, sorry!
Let me clarify:

- My server
- My Desktop (fixed computer in my room)
- Two laptops

I see, and between which computers are the ones giving you poor iperf results?

 

[Proxipupuce]

Can I get a clarification, which computer is which? The terms desktop and PC are kind of confusing me. You have a server, a desktop, and two laptops?

Also, which ones are giving you poor iperf results?

The bad results are :

- My server to the desktop in my room and the first laptop
- The desktop in my room to the server and the two laptops
These results in both directions.

The good results are :

- My server to the second laptop, both ways.

 

[Whattteva]

The bad results are :

- My server to the desktop in my room and the first laptop
- The desktop in my room to the server and the two laptops
These results in both directions.

So let me clarify this:

• Server -> desktop and desktop -> server
• Server -> laptop1 and laptop1 -> server

I'm guessing laptop2 is probably pretty old and maybe connecting through 802.11n instead of 802.11ac possibly or maybe just a poor WiFi card.
Did you do any testing between your desktop and laptop1 and also between laptop1 and laptop2?

 

[Proxipupuce]

So let me clarify this:

• Server -> desktop and desktop -> server
• Server -> laptop1 and laptop1 -> server

Desktop -> Server and Server -> Desktop: <100 mbps
laptop1 -> Server and Server -> laptop1 : <100 mbps
laptop2 -> Server and Server -> laptop2 : >250 mbps

Did you do any testing between your desktop and laptop1 ?

Desktop -> laptop1 & laptop 2 : <100 mbps

between laptop1 and laptop2?

laptop1 -> laptop 2 : around 100 mbps

 

[Whattteva]

Desktop -> Server and Server -> Desktop: <100 mbps
laptop1 -> Server and Server -> laptop1 : <100 mbps
laptop2 -> Server and Server -> laptop2 : >250 mbps

Desktop -> laptop1 & laptop 2 : <100 mbps

laptop1 -> laptop 2 : around 100 mbps

Looking at those results. It looks to me that your desktop and laptop1 are the slow links. Do you by any chance know the WiFi card of laptop1 or maybe make/model of laptop1?

 

[Proxipupuce]

Looking at those results. It looks to me that your desktop and laptop1 are the slow links. Do you by any chance know the WiFi card of laptop1 or maybe make/model of laptop1?

Laptop1 : Qualcomm Atheros QCA61x4A Wirelles Network Adaptater

Desktop : MSI b450M mortar MAX with realtek pcie gbe family controller. With this, I use an Asus RT-AX55.

 

[Proxipupuce]

Just in case, I looked into buying a PCIe network card:

- Intel I210-T1: 15$.
- Intel Gigabit CT Desktop: 10$.

For so little money, it will only be beneficial.
Which one do you prefer between the two?

I saw that the Intel Gigabit CT Desktok was old, in PCI 1.0 from 2008, while the Intel I210-T1 was in 2.0 from 2013 and still referenced by Intel.

 

[Whattteva]

Laptop1 : Qualcomm Atheros QCA61x4A Wirelles Network Adaptater

Desktop : MSI b450M mortar MAX with realtek pcie gbe family controller. With this, I use an Asus RT-AX55.

Looks to me that your desktop and laptop1 are the weak links. It will be nice if you could test against an ethernet NIC that is actually known to reach close to 1G.

I saw that the Intel Gigabit CT Desktok was old, in PCI 1.0 from 2008, while the Intel I210-T1 was in 2.0 from 2013 and still referenced by Intel.

Yeah, I would prefer the i210 since it's newer. Careful with these used Intel cards. Some of them can be counterfeit though I think that's more common with the multi-port cards.

 

[Proxipupuce]

Looks to me that your desktop and laptop1 are the weak links. It will be nice if you could test against an ethernet NIC that is actually known to reach close to 1G.

Unfortunately, I don't have a network card to test. I'll have to wait until I get one.
However, I would like to know: How come my connection to the server is so bad, if I can download games or movies at 400mbps?

I would prefer the i210 since it's newer. Careful with these used Intel cards. Some of them can be counterfeit though I think that's more common with the multi-port cards.

Indeed, of the 4 cards I saw, only one is an original, with the labels and the certificate. The others are counterfeit.
So I'm thinking of going with an Intel Gigabit CT Desktop, because I'm sure it's a real Intel card.

 

[Whattteva]

Unfortunately, I don't have a network card to test. I'll have to wait until I get one.
However, I would like to know: How come my connection to the server is so bad, if I can download games or movies at 400mbps?

Yeah, that's a head scratcher. You should never see local speeds < internet speeds. Usually your internet speeds will be less unless you're one of those lucky ones that have a 1G internet connection. Mine, for example is only 350 Mbps max and that's only during off-peak times. Realistically, my internet speed is more like 300M.