The close lightning bolt induces saturation current in the LAN cabling. CAT5 cabling is unshielded. The differential signal elimates noise problems as long as noise signals are within certain specs. A lightning bolt or an EMP far exceeds the few volt limit and saturates the clamping protection diodes on the circuitries input stages. These gates are then permanently shorted.
Device is now useless.
The close lightning bolt induces saturation current in theI have NEVER seen an event on any LAN I was on that I could correlate with nearby lighting discharges; I HAVE witnessed latch state change on equipment from nearby strikes, equipment that was poorly shielded/open owing to the developmental state of said equipment ...I still think you had a 'mains' event that probably 'dropped out' the 'switching' power supply on a node/switch somewhere but you're blaming it on the transport media itself ...
Device is now useless.A 'clue' that you probably had a 'mains' arc across the power supply; not uncommon. (Related failure inducements include a phenom called 'ground bounce')A main's arc is not the same thing as EMP as an arc is actual electron flow whereas EMP is only a field until it induces a current in a conductor.
APC UPS systems have a Phone/Modem jack that is optically isolated along with the power supply so that doesn't happen.