Yes I would also like to add some points... "Attenuation" is what happens when the intensity of light decreases as it travels through an optical fiber due to glass not being 100% transmissive, and absorbing some light. It is not caused by short distances. If you are using 10km SFP+ 1310nm which seems to be one of the common ones there is no risk of burning, the power isn't very high, I can't speak for 40km transceivers or >10gbit. Of course it will not work with the multimode transceiver because the wavelength is different, and it could damage the optic, but "short distance" between two identical transceivers was never a problem in my experience with them. They do run toasty though. SFPs are really cheap (10km SM ones can be found for $10 on ebay) if your hardware does not require a certain brand. If it does, aside from flashing the SFP firmware to another manufacturer (which may be more expensive than just buying their own SFP), have fun with the cost. Even the bare fiber is not that expensive ($100-$200 a km?). The expensive bit is actually laying the fiber across buildings and the networking equipment on either end. I am not speaking from a ISP perspective here, of course everything will be 100x the cost if you are an ISP or a customer of an ISP, because that is how businesses work. Next, the speedtest servers are usually 10gbit if they go over 1000. The limitation is the speedtest client itself and your web browser (or even your CPU). The best way to test in a "real world scenario" this is definitely bittorrent, find a large swarm and try downloading a file. Or you can use wget with test files or any public https://iperf.fr/iperf-servers.php server. May be hard to find ones in canada though that have 10gbit, it is a lot more common in europe.