Unfortunately it is rather false-positive prone when applied to fences, significant wind loading can cause the chain-link to lash back and forth and repeatedly deflect the cable. The resulting alarm fatigue was implicated in the somewhat (in)famous case of the nuns as site Y-12. Nonetheless this application is very common at federal installations and also certain non-federal infrastructure installations, e.g. the legacy iron fences around Portland, Oregon's surface water reservoirs (where remote-controlled PTZ cameras appear to be used for rapid remote confirmation of FiberPatrol detections... you can give the fence a shake and watch a camera pan to you almost immediately).
This is what I was told by the person who saw the demo. I believe them, because it'd just be a matter of sending an impulse (or sweeping a tone) and looking at all of the spikes in the impulse response.
I've seen this info in the Realtek NIC diagnostic tool for windows (on one machine, anyway), and I've got a couple of semi-managed switches that show a line length estimate which I assume is based on similar measurements.
At least some gigabit ethernet interfaces have built in weaker TDR that measures the length of each twisted pair, but I've only ever had one computer that showed that info in the BIOS/UEFI.
So, maybe if TDR is used for inspection, we can rest assured.