Question: If you assume that at some point (until it becomes ubiquitous), the mesh network must connect to existing broadband providers for bandwidth, how could you avoid charging?
I'm not suggesting profiting off of users, but rather covering the bandwidth costs incurred by each user in the network. In theory, the costs approach zero over time (once density is reached in each distribution of nodes), but surely some monetary transfer is ok.
Aether is not a full-fledged mesh network in that sense, it still relies on an existing IP connection between the nodes. But in cases of full blown mesh networks, there needs to be 'exits' from the mesh to the normal Internet, and those points would cost money to run. In that case, though, those exits would be clearly identified, i.e. not fleeting interstitial nodes on the mesh network, and you could direct payment to them. The problem arises when you try to transfer money from a mesh node to another, not from a mesh node to a known location (the exit).
Agreed. I'm approaching this thinking that the exit points can be backbone providers and we could build simple software to enable users to pay for the bandwidth (to the exit) that they've utilized. Connectivity within the network would be free and we could even place peering points at the exit to reduce outbound bandwidth.
Then it just becomes a question of incentivizing users by consistently driving down costs (in theory) to promote density of nodes.
Yes, this could be an intermediary state between a full–blown mesh network which everything is fully peered, and a fully centralised network such as today's Internet. It all depends on everybody carrying their own weight, though, and of all, that seems to be the farthest away concept so far. Phones and other mobile devices are consuming an increasingly sizeable chunk of the pool and those devices are barely able to connect and to their work, much less work for others due to battery restrictions, which are ultimately bound by the laws of physics.
This is also the reason Aether will most probably never have a mobile version. Not until they finally make nuclear fusion work, and compact it enough to make batteries with it!
I'm not suggesting profiting off of users, but rather covering the bandwidth costs incurred by each user in the network. In theory, the costs approach zero over time (once density is reached in each distribution of nodes), but surely some monetary transfer is ok.