The idea of sending media before you actually have a destination address is interesting because the patent mentioned network hops. If that just means storing it then that's boring, but if it meant designing a live network that could update and rewrite destinations in flight then that is interesting. In the real time media systems I know of you always know the endpoint address before sending, even if that address is merely the next hop or a bridge.
ethernet has a broadcast mode in which the specific destination is not known (obviously "all nodes on some range" but that's not specific), and media is one of the tentpole use-cases of these sorts of multicast
