*UPDATE
1/12/2021 – Latency Testing Added
One of a limitation of standard SFP to ethernet transceiver module when compared to direct Ethernet port is the length of coverage. In fact, I have tried connect my iMac using 10G adapter to switch via SFP+ port with standard 10G Ethernet module and that did not work despite the distance between two is right around the limit (27m/90ft). So I’ve decided to order one of the transceiver module that is rated for 80m/260ft coverage.
Improve length
Approximately 27m/90ft between my switch (Unifi Aggregation Switch) and iMac 2020 with 10Gbps Ethernet to thunderbolt 3 adapter, the new transceiver achieved 10G performance both up and down with iPerf speed test. Originally, I had some fluctuation in download speed, but after plug and unplug transceiver and ethernet cable, I got full 9.4 Gbps both up and down.
According to official FS site user comments, some are reporting 40m and 50m 10G speed with Cat 7 cables. Since Cat 6 cable should be covering up to 55m 10G speed, I am fairly certain Cat 6 would have done the job in above cases as well with this transceiver.
Latency
Using transceiver means another electrical module. So how much latency will this create? Here is a quick, perhaps crude test using classic “ping” command.
Setup
The client device I used is iMac 2020. Destination are either router (UDM Pro) or NAS server.
There are two potential pathways were evaluated.
FS Transceiver was used in 10G Pathway where iMac 2020 was connected to USW-Aggregation. Both UDM Pro and NAS were also connected to USW Aggregation Switch either via DAC (UDM Pro) or ethernet to transceiver (NAS). Of note, iMac end used ethernet to thunderbolt 3 adapter for 10G connection.
In 1G pathway, iMac was connected to USW Pro 24 Gen 2 directly by cat 6 ethernet cable. There is also 1G direct ethernet connection between the switch to NAS. So this pathway has no transceiver. Connection to UDM Pro here require Gen 2 switch to go through USW Aggregation switch with two being connected by DAC.
Testing Method
I have run around 20 (17-21) “ping” command from iMac to the destination.
Result
Discussion
The result interpretation above is simple. When we look at the average, there just isn’t enough difference amongst tested conditions. One may look at first and think 10G pathway to the NAS is slower than the other 3 condition; however, when I re-run the 10G pathway to LAN, it became 0.169/0.241/0.349/0.040. In another word, I think above numbers are just margins of errors when I test.
So what this pertains to for FS transceiver module is that it does not add any additional latency to the network. As a side note, the same can be said to DAC module because 1G to router pathway has two DACs while 1G path to NAS has a single ethernet cable in place of 2 DACs (plus USW-Aggregation switch) yet two have essentially the same latency.
No N-BaseT
One unfortunate part is that this module do not add a capability of 2.5G or 5G N-BaseT support to my UniFi aggregation switch. Although the switch itself is only rated for 10Gbps, I was hoping this transceiver module can make those non-N-BaseT SFP+ ports into N-BaseT ports in addition to 10GbE port, but this did not work for me.
Conclusion
Long range 10G transceiver serves its main purpose i.e. 10Gbps data transfer rate beyond 30m limit. At the cost of almost 2x the regular 10G SFP+ to RJ45 transceiver module, I had hoped these unit provide N-BaseT support, but perhaps I am asking a bit too much here.