Are there 2.5GBase-T SFP+ modules?

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

bbqdt

Member
Sep 15, 2019
93
64
18
Ordered 2 of the Optcore version! Seems like it should work since it links with the switch at 10G and converts internally.

Going to try with a genuine Cisco AIR-PWRINJ6. Seen some reports that NBase-T will link at 2.5 or 5 through a 1Gb injector, just not 10.
@mb300sd Any update? did they work as expected?
 

ERDrPC

Member
Aug 14, 2012
36
3
8
Just a nit-pic - 10GbaseT does support PoE. For example Ubiquiti has a small 10GbaseT PoE switch on the market.
Would it make sense to use this to power the new ruckus 5Gbase-T APs x 3. Have found it nearly impossible to find 5.0 5Gbase-T poe++ mini switch for this purpose. This seems to be the first one. I have a mikrotik router and multiple switches. I just need the sfp+ uplink to my mikrotik. Can all of the unifi special juice be turned off so that it acts as a go-between the router and APs?
 

keenanj

New Member
May 8, 2020
7
2
3
Would it make sense to use this to power the new ruckus 5Gbase-T APs x 3. Have found it nearly impossible to find 5.0 5Gbase-T poe++ mini switch for this purpose. This seems to be the first one. I have a mikrotik router and multiple switches. I just need the sfp+ uplink to my mikrotik. Can all of the unifi special juice be turned off so that it acts as a go-between the router and APs?
Yes I am running the R750 on the ECS2512FP and it is working great these switches only do 2.5Gb but that is more than enough speed for any Wi-Fi 6 AP.
 

keenanj

New Member
May 8, 2020
7
2
3
[
Would it make sense to use this to power the new ruckus 5Gbase-T APs x 3. Have found it nearly impossible to find 5.0 5Gbase-T poe++ mini switch for this purpose. This seems to be the first one. I have a mikrotik router and multiple switches. I just need the sfp+ uplink to my mikrotik. Can all of the unifi special juice be turned off so that it acts as a go-between the router and APs?
The Ubiquiti switch is weak only 4 ports unknown PoE budget (spec says 210w PS with 60W per port does not add up) no cloud management no support and typical UBNT bad quality.
 

madbrain

Active Member
Jan 5, 2019
212
44
28
Guys I just tested the Wiitek SFP+ to RJ45 Copper module with the EnGenius ECS2512FP Multi-Gigabit 2.5GbE PoE++ switch and it is working well I did a blog post here Convert a SFP+ Port To Copper Multi-Gig 2.5/5/10Gbps with a 10Gbase-T Adapter – Keenan Systems Wi-Fi
Curious - what kind of GUI tool were you using in your test of the Cable Creation 2.5 Gbps NIC ?
Did you test both directions with iperf ? And were you using jumbo frames or not.
I have had problems with this NIC with a Trendnet NBASE-T switch.

 

keenanj

New Member
May 8, 2020
7
2
3
Curious - what kind of GUI tool were you using in your test of the Cable Creation 2.5 Gbps NIC ?
Did you test both directions with iperf ? And were you using jumbo frames or not.
I have had problems with this NIC with a Trendnet NBASE-T switch.

I use jperf it is a graphical front end to iperf enabling jumbos at 4088b on the cable creations 2.5G adapter driver 10.3920518 resulted in full speed with the ECS2512FP set to 9014b and my 10g workstation set to 4088b same speed in both directions! The ECS2512FP is working great with the cable creations USB adapter.
 

Attachments

Last edited:
  • Like
Reactions: madbrain

bbqdt

Member
Sep 15, 2019
93
64
18
Just to follow up on this thread, I obtained a couple of Super-micro/Aquantia branded AQS-107 SFP+ Transceivers on eBay for around 10 bucks and they do link at 2.5g in an icx 6610 to a cheap usb 3.1 Realtek adapter.

The icx switch shows them linking at 10g and they just only transmit up to 2.5g worth or traffic.

They also work fine at 10g, but that is expected.

I can’t speak for the opticore ones, but it stands to reason that they would work too.
 
Last edited:

madbrain

Active Member
Jan 5, 2019
212
44
28
Just to follow up on this thread, I obtained a couple of Super-micro/Aquantia branded AQS-107 SFP+ Transceivers on eBay for around 10 bucks and they do link at 2.5g in an icx 6610 to a cheap usb 3.1 Realtek adapter.

The icx switch shows them linking at 10g and they just only transmit up to 2.5g worth or traffic.

They also work fine at 10g, but that is expected.

I can’t speak for the opticore ones, but it stands to reason that they would work too.
Have you run iperf3 in both directions ? Many of us are having issues when mixing speeds. Even on NBASE-T switches (no SFP+ involved).

 

bbqdt

Member
Sep 15, 2019
93
64
18
Server (40gbe) to client (2.5gbe) -

Code:
# iperf3 -t 45 -c 10.0.0.190
Connecting to host 10.0.0.190, port 5201
[  4] local 10.0.0.249 port 36134 connected to 10.0.0.190 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   283 MBytes  2.37 Gbits/sec    0    277 KBytes    
[  4]   1.00-2.00   sec   280 MBytes  2.35 Gbits/sec    0    277 KBytes    
[  4]   2.00-3.00   sec   281 MBytes  2.36 Gbits/sec    0    274 KBytes    
[  4]   3.00-4.00   sec   281 MBytes  2.36 Gbits/sec    0    277 KBytes    
[  4]   4.00-5.00   sec   280 MBytes  2.35 Gbits/sec    1    274 KBytes    
[  4]   5.00-6.00   sec   281 MBytes  2.36 Gbits/sec    0    277 KBytes    
[  4]   6.00-7.00   sec   280 MBytes  2.35 Gbits/sec    6    274 KBytes    
[  4]   7.00-8.00   sec   280 MBytes  2.35 Gbits/sec    0    274 KBytes    
[  4]   8.00-9.00   sec   281 MBytes  2.36 Gbits/sec    8    280 KBytes    
[  4]   9.00-10.00  sec   270 MBytes  2.26 Gbits/sec    0   5.66 KBytes    
[  4]  10.00-11.00  sec   229 MBytes  1.92 Gbits/sec    8    243 KBytes    
[  4]  11.00-12.00  sec   280 MBytes  2.34 Gbits/sec   14    249 KBytes    
[  4]  12.00-13.00  sec   281 MBytes  2.35 Gbits/sec   25    243 KBytes    
[  4]  13.00-14.00  sec   281 MBytes  2.35 Gbits/sec    0    243 KBytes    
[  4]  14.00-15.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes    
[  4]  15.00-16.00  sec   281 MBytes  2.36 Gbits/sec    0    246 KBytes    
[  4]  16.00-17.00  sec   282 MBytes  2.36 Gbits/sec    0    246 KBytes    
[  4]  17.00-18.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  18.00-19.00  sec   280 MBytes  2.35 Gbits/sec    0    255 KBytes    
[  4]  19.00-20.00  sec   281 MBytes  2.36 Gbits/sec   23    246 KBytes    
[  4]  20.00-21.00  sec   186 MBytes  1.56 Gbits/sec   11    243 KBytes    
[  4]  21.00-22.00  sec   281 MBytes  2.36 Gbits/sec    1    243 KBytes    
[  4]  22.00-23.00  sec   281 MBytes  2.36 Gbits/sec    0    240 KBytes    
[  4]  23.00-24.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes    
[  4]  24.00-25.00  sec   281 MBytes  2.35 Gbits/sec    0    246 KBytes    
[  4]  25.00-26.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  26.00-27.00  sec   281 MBytes  2.35 Gbits/sec    0    243 KBytes    
[  4]  27.00-28.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  28.00-29.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  29.00-30.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  30.00-31.00  sec   206 MBytes  1.73 Gbits/sec    0    243 KBytes    
[  4]  31.00-32.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  32.00-33.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes    
[  4]  33.00-34.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  34.00-35.00  sec   281 MBytes  2.36 Gbits/sec    0    240 KBytes    
[  4]  35.00-36.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes    
[  4]  36.00-37.00  sec   281 MBytes  2.36 Gbits/sec    4    246 KBytes    
[  4]  37.00-38.00  sec   281 MBytes  2.35 Gbits/sec    0    246 KBytes    
[  4]  38.00-39.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
[  4]  39.00-40.00  sec   281 MBytes  2.35 Gbits/sec    0    243 KBytes    
[  4]  40.00-41.00  sec   226 MBytes  1.90 Gbits/sec    0   5.66 KBytes    
[  4]  41.00-42.00  sec   274 MBytes  2.30 Gbits/sec    0    243 KBytes    
[  4]  42.00-43.00  sec   279 MBytes  2.34 Gbits/sec   10    240 KBytes    
[  4]  43.00-44.00  sec   274 MBytes  2.30 Gbits/sec  230    243 KBytes    
[  4]  44.00-45.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes    
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-45.00  sec  12.0 GBytes  2.30 Gbits/sec  341             sender
[  4]   0.00-45.00  sec  12.0 GBytes  2.30 Gbits/sec                  receiver

It does show some dropped packets. Client to server has about the same error rate.

I only have a 2.5g adapter. The problem in that thread might only show up if I had a 5g adapter.
 
Last edited:

bbqdt

Member
Sep 15, 2019
93
64
18
Contrary to what the web guide might suggest, the ICX 6610 does support rate limiting on physical ports, the following does work on my switch -

Code:
SSH@ICX6610-24P Router(config)#interface ethernet 1/3/5
SSH@ICX6610-24P Router(config-if-e10000-1/3/5)#rate-limit output shaping 2110000
Outbound Rate Shaping on Port 1/3/5 Config: 2110000 Kbps, Actual: 2109375 Kbps
That will limit the output rate of the port to around 2gbit/s. That seems to reduce my dropped packet count to around 0.

Maybe try that with your 5gbit adapters.

edit - just noticed someone said they tried this in the other thread, I’ll leave this info here though
 
Last edited:

madbrain

Active Member
Jan 5, 2019
212
44
28
Server to client (connected via 2.5g) -

Code:
# iperf3 -t 45 -c 10.0.0.190
Connecting to host 10.0.0.190, port 5201
[  4] local 10.0.0.249 port 36134 connected to 10.0.0.190 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   283 MBytes  2.37 Gbits/sec    0    277 KBytes     
[  4]   1.00-2.00   sec   280 MBytes  2.35 Gbits/sec    0    277 KBytes     
[  4]   2.00-3.00   sec   281 MBytes  2.36 Gbits/sec    0    274 KBytes     
[  4]   3.00-4.00   sec   281 MBytes  2.36 Gbits/sec    0    277 KBytes     
[  4]   4.00-5.00   sec   280 MBytes  2.35 Gbits/sec    1    274 KBytes     
[  4]   5.00-6.00   sec   281 MBytes  2.36 Gbits/sec    0    277 KBytes     
[  4]   6.00-7.00   sec   280 MBytes  2.35 Gbits/sec    6    274 KBytes     
[  4]   7.00-8.00   sec   280 MBytes  2.35 Gbits/sec    0    274 KBytes     
[  4]   8.00-9.00   sec   281 MBytes  2.36 Gbits/sec    8    280 KBytes     
[  4]   9.00-10.00  sec   270 MBytes  2.26 Gbits/sec    0   5.66 KBytes     
[  4]  10.00-11.00  sec   229 MBytes  1.92 Gbits/sec    8    243 KBytes     
[  4]  11.00-12.00  sec   280 MBytes  2.34 Gbits/sec   14    249 KBytes     
[  4]  12.00-13.00  sec   281 MBytes  2.35 Gbits/sec   25    243 KBytes     
[  4]  13.00-14.00  sec   281 MBytes  2.35 Gbits/sec    0    243 KBytes     
[  4]  14.00-15.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes     
[  4]  15.00-16.00  sec   281 MBytes  2.36 Gbits/sec    0    246 KBytes     
[  4]  16.00-17.00  sec   282 MBytes  2.36 Gbits/sec    0    246 KBytes     
[  4]  17.00-18.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  18.00-19.00  sec   280 MBytes  2.35 Gbits/sec    0    255 KBytes     
[  4]  19.00-20.00  sec   281 MBytes  2.36 Gbits/sec   23    246 KBytes     
[  4]  20.00-21.00  sec   186 MBytes  1.56 Gbits/sec   11    243 KBytes     
[  4]  21.00-22.00  sec   281 MBytes  2.36 Gbits/sec    1    243 KBytes     
[  4]  22.00-23.00  sec   281 MBytes  2.36 Gbits/sec    0    240 KBytes     
[  4]  23.00-24.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes     
[  4]  24.00-25.00  sec   281 MBytes  2.35 Gbits/sec    0    246 KBytes     
[  4]  25.00-26.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  26.00-27.00  sec   281 MBytes  2.35 Gbits/sec    0    243 KBytes     
[  4]  27.00-28.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  28.00-29.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  29.00-30.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  30.00-31.00  sec   206 MBytes  1.73 Gbits/sec    0    243 KBytes     
[  4]  31.00-32.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  32.00-33.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes     
[  4]  33.00-34.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  34.00-35.00  sec   281 MBytes  2.36 Gbits/sec    0    240 KBytes     
[  4]  35.00-36.00  sec   280 MBytes  2.35 Gbits/sec    0    243 KBytes     
[  4]  36.00-37.00  sec   281 MBytes  2.36 Gbits/sec    4    246 KBytes     
[  4]  37.00-38.00  sec   281 MBytes  2.35 Gbits/sec    0    246 KBytes     
[  4]  38.00-39.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
[  4]  39.00-40.00  sec   281 MBytes  2.35 Gbits/sec    0    243 KBytes     
[  4]  40.00-41.00  sec   226 MBytes  1.90 Gbits/sec    0   5.66 KBytes     
[  4]  41.00-42.00  sec   274 MBytes  2.30 Gbits/sec    0    243 KBytes     
[  4]  42.00-43.00  sec   279 MBytes  2.34 Gbits/sec   10    240 KBytes     
[  4]  43.00-44.00  sec   274 MBytes  2.30 Gbits/sec  230    243 KBytes     
[  4]  44.00-45.00  sec   281 MBytes  2.36 Gbits/sec    0    243 KBytes     
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-45.00  sec  12.0 GBytes  2.30 Gbits/sec  341             sender
[  4]   0.00-45.00  sec  12.0 GBytes  2.30 Gbits/sec                  receiver

It does show some dropped packets. But client to server has about the same error rate.

I only have a 2.5g adapter. The problem in that thread might only show up if I had a 5g adapter.
Did you try in the other direction ? All the problems observed were when the slower client was receiving (-R switch in iperf3). By default, the client in iperf3 is the sender.

The problem isn't specific to 5G. I saw the same problem with 2.5G client against 10G server, and 5G client against 10G server.
Didn't try 2.5G client against 5G server.
 

bbqdt

Member
Sep 15, 2019
93
64
18
Did you try in the other direction ? All the problems observed were when the slower client was receiving (-R switch in iperf3). By default, the client in iperf3 is the sender.

The problem isn't specific to 5G. I saw the same problem with 2.5G client against 10G server, and 5G client against 10G server.
Didn't try 2.5G client against 5G server.
That is the faster server (40gbe to switch) sending to 2.5g client. 10.0.0.190 is the client. Same result with -R on client.
 
  • Like
Reactions: lowfat

Biomesh

New Member
Aug 15, 2020
4
1
3
For those having performance issues, do you get the same receive speeds even when linked at 1Gbps? For me this is the case. I have tried with pause frames enabled and disabled with no difference. This was tested with two Mikrotik switches - crs326 and crs309.
I would like to know if this is more of an sfp+ module issue or an issue with the switch. The module in the switch always shows linked at 10G but the client adapter shows the correct speed too. I have a feeling this is the issue.
 

lowfat

Active Member
Nov 25, 2016
131
91
28
40
For those having performance issues, do you get the same receive speeds even when linked at 1Gbps? For me this is the case. I have tried with pause frames enabled and disabled with no difference. This was tested with two Mikrotik switches - crs326 and crs309.
I would like to know if this is more of an sfp+ module issue or an issue with the switch. The module in the switch always shows linked at 10G but the client adapter shows the correct speed too. I have a feeling this is the issue.
It happened for me even w/ a 1GbE Intel NIC on a CRS305 w/ IPolex transceiver. It also showed 10G link on the switch but the NIC showed the correct link.
 

Biomesh

New Member
Aug 15, 2020
4
1
3
Well, looking at some comments earlier in this thread made by bbqdt, I set the egress rate on my RouterOS switch port (switch->port) to 2400M for a 2.5 G connection. This made it possible to get full speed (with low retransmits) on my connection. The issue was basically the switch spamming the sfp module causing the issues. By controlling the speed at the switch level, you should be able to have full operation. If you are using SWos, you will have to do testing to see how to limit the egress rate.
 
  • Like
Reactions: lowfat

lowfat

Active Member
Nov 25, 2016
131
91
28
40
Well, looking at some comments earlier in this thread made by bbqdt, I set the egress rate on my RouterOS switch port (switch->port) to 2400M for a 2.5 G connection. This made it possible to get full speed (with low retransmits) on my connection. The issue was basically the switch spamming the sfp module causing the issues. By controlling the speed at the switch level, you should be able to have full operation. If you are using SWos, you will have to do testing to see how to limit the egress rate.
swOS has no engress options AFAICT. But I played w/ it in RouterOS. Setting it to 2400M has zero impact for me. Still about 20k retries w/ a max receiver throughput of about 1.3Mbps. But if I set it to 1300M, retries dropped to 2K.
 

Biomesh

New Member
Aug 15, 2020
4
1
3
I didn't test the acl functionality of swos, but from what I can see that would be the only way. That might be ingress though.

I am using an ipolex 10gbase-t sfp+ adapter connected to a mikrotik crs309 running ros 6.47.1. I am connecting to a rtl-8156 usb 3.0 adapter(sabrent) and I also tested with an integrated 1G Intel nic on my laptop.

I tested the qsfptek adapter and had the same results. I ended up returning it, but it seems the issue was just the switch config in my case.

I can post my switch config if it helps.
 

lowfat

Active Member
Nov 25, 2016
131
91
28
40
I didn't test the acl functionality of swos, but from what I can see that would be the only way. That might be ingress though.

I am using an ipolex 10gbase-t sfp+ adapter connected to a mikrotik crs309 running ros 6.47.1. I am connecting to a rtl-8156 usb 3.0 adapter(sabrent) and I also tested with an integrated 1G Intel nic on my laptop.

I tested the qsfptek adapter and had the same results. I ended up returning it, but it seems the issue was just the switch config in my case.

I can post my switch config if it helps.
Seems your setup should be almost the same as mine. What version of the Realtek drivers are you using? Jumbo frames? Any other notable changes? Tried various settings for hours this morning and I just can't get more than 1Gb consistently for RX no matter what I'm trying.