F5F Stay Refreshed Power Users Networks 10gig Bottleneck

10gig Bottleneck

10gig Bottleneck

Pages (2): Previous 1 2
N
nuclernoah101
Member
120
01-18-2026, 06:35 AM
#11
N
nuclernoah101
01-18-2026, 06:35 AM #11

A
AubreyJJJ
Junior Member
9
01-18-2026, 10:10 AM
#12
Still under the 10gbit mark (7–8gbit), it’s worth getting a much stronger option than standard gigabit. This should give you around 1GB/s in raw data transfer, possibly due to disk limitations others discussed.
A
AubreyJJJ
01-18-2026, 10:10 AM #12

Still under the 10gbit mark (7–8gbit), it’s worth getting a much stronger option than standard gigabit. This should give you around 1GB/s in raw data transfer, possibly due to disk limitations others discussed.

M
ManMallow
Member
223
01-18-2026, 02:46 PM
#13
You tested Iperf on two workstations using the same network connection. The results showed speeds around 3-4 Gbps, which is significantly higher than the advertised 10 Gbps. This suggests potential issues with the switch or network configuration.
M
ManMallow
01-18-2026, 02:46 PM #13

You tested Iperf on two workstations using the same network connection. The results showed speeds around 3-4 Gbps, which is significantly higher than the advertised 10 Gbps. This suggests potential issues with the switch or network configuration.

K
koopa14
Member
114
01-25-2026, 07:32 AM
#14
The speed you referenced aligns well with the performance seen in your initial PC1 to PC2 transfer (3~4 Gbit matching the 377 MB/s you recorded). It’s unclear whether the issue is with the switch or your computers.
K
koopa14
01-25-2026, 07:32 AM #14

The speed you referenced aligns well with the performance seen in your initial PC1 to PC2 transfer (3~4 Gbit matching the 377 MB/s you recorded). It’s unclear whether the issue is with the switch or your computers.

A
Aye_Aye
Junior Member
9
01-25-2026, 09:26 AM
#15
Absolutely, I see. I’m not sure what other adjustments are possible with the Switch, but it seems the PC is already configured properly for the 10Gbit card.
A
Aye_Aye
01-25-2026, 09:26 AM #15

Absolutely, I see. I’m not sure what other adjustments are possible with the Switch, but it seems the PC is already configured properly for the 10Gbit card.

D
ditox95
Member
234
01-25-2026, 01:29 PM
#16
I think there might be a misunderstanding. At first I wondered if the NIC on your PC was the issue, since many modern 10G Ethernet cards are from Aquantia. However, it seems you have a genuine Intel product, so that’s likely not the cause. I checked two servers in your basement connected via the same Cisco switch: one showed a stable 10Gbps transfer, and the other also confirmed similar speeds. I’m seeing around 10Gbps consistently, even without jumbo frames enabled. From my Mac in the office to a server in the basement means it’s passing through two switches, which could affect performance. I’m focusing on the switch itself.
D
ditox95
01-25-2026, 01:29 PM #16

I think there might be a misunderstanding. At first I wondered if the NIC on your PC was the issue, since many modern 10G Ethernet cards are from Aquantia. However, it seems you have a genuine Intel product, so that’s likely not the cause. I checked two servers in your basement connected via the same Cisco switch: one showed a stable 10Gbps transfer, and the other also confirmed similar speeds. I’m seeing around 10Gbps consistently, even without jumbo frames enabled. From my Mac in the office to a server in the basement means it’s passing through two switches, which could affect performance. I’m focusing on the switch itself.

A
aman_da
Junior Member
6
01-25-2026, 03:29 PM
#17
What's wrong with Aquantia?
A
aman_da
01-25-2026, 03:29 PM #17

What's wrong with Aquantia?

F
Fa837241
Member
100
01-27-2026, 09:38 AM
#18
Looking for issues but can't spot anything obvious. You seem to be mainly used to Cisco switches. What configurations are you using so I can compare? The server is running on 4 ports with a frame size of 1500, and flow control is off.
F
Fa837241
01-27-2026, 09:38 AM #18

Looking for issues but can't spot anything obvious. You seem to be mainly used to Cisco switches. What configurations are you using so I can compare? The server is running on 4 ports with a frame size of 1500, and flow control is off.

E
Error_Sans55
Member
245
01-27-2026, 11:12 AM
#19
Another concept came to mind: could we link one device directly to another without using a switch? For example, connecting workstation 1 straight to workstation 2 or any workstation to the Synology NAS. Running iperf this way would help identify issues in the middle device, likely the switch itself.
E
Error_Sans55
01-27-2026, 11:12 AM #19

Another concept came to mind: could we link one device directly to another without using a switch? For example, connecting workstation 1 straight to workstation 2 or any workstation to the Synology NAS. Running iperf this way would help identify issues in the middle device, likely the switch itself.

C
Commando__
Senior Member
744
01-27-2026, 08:15 PM
#20
I don’t have any relevant settings, and I’m not comfortable with Netgear switches. My setup focuses on LAGs and VLANs, and I’m not concerned about impacting a single 10Gbit/sec stream.
C
Commando__
01-27-2026, 08:15 PM #20

I don’t have any relevant settings, and I’m not comfortable with Netgear switches. My setup focuses on LAGs and VLANs, and I’m not concerned about impacting a single 10Gbit/sec stream.

Pages (2): Previous 1 2