I am speaking strictly layer 2. If a naive switch were to forward a frame with FF:FF:FF:FF:FF:FF as its source mac and a even more naive client would respond to that frame, this time with the broadcast address in the DST, it could lead to some undesired results. If the original frame also had the DST set to broadcast and we had multiple of these naive clients, all hell would break lose. This is why I think switch implementations simply drop frames with multi- or broadcast as their SRC.
But what I want to know is if there is a specific standard like a RFC that specifies this? Or is this just in the discression (and maybe best interest) of the switch implementation?
No comments:
Post a Comment