Because this facilitates the kind of bespoke communication infrastructure with which SQRL could prove synergetic, this proposal carries, for me, a special urgency:
Peer-to-peer, aka any-source, multi-cast still doesn’t scale well, but this is a protocol for single-source which is then propagated by the network as a multi-cast in an intelligently managed fashion brokered by the end-points, in-app.
Credit to Esther Payne for surfacing this on Twitter:
Nice find @philodygmn! The inefficiency of unicast at scale is simply mind-boggling when you think about it in context next to a multicast solution. I’ve always thought implementing multicast is a problem that can (and should, unlike many things caught up in the software-defined trend) be solved via software. Hope this effort attracts the correct talent.
I should note explicitly the implication of end-point administration: the transport layer remains traffic-agnostic aka net neutral and perhaps just as importantly free of any need to integrate with content distributors, which the network infrastructure folks should be elated about. It means that transport hardware need not be colonized by content creators, and that can easily be seen to have profound beneficial and preservative political (to say nothing of competitive) effect.