Quote from: Vector on September 05, 2011, 10:18:47 AM

I'm not really in a position to care to argue for this (dropped the project months ago), but for the sake of discussion:

4.) ISP that state they are not allowed to resell the connection

This isn't about reselling connections - it would require licensing unused wireless bands (most likely wiMAX frequency areas) which is difficult, and one of the reasons I'm not pursuing the idea.

During the last spectrum auction in Canada, the incumbent cell phone carriers spent about $4 billion CAD, doubling their amortized infrastructure costs in the process. The cell phone carriers expect to recouperate that cost by charging the average cell phone user $40/month. As I said, spectrum is a shared resource. You proposal seems to assume only a handful of users with everybody else acting as relays.

Quote from: Elwar

Each node should not care whether the packet is coming from a wireless connection or a wire connection or even a connection through the Internet.

With a shared medium like air, it is easier to get routing loops.

Quote from: Elwar

I think if each node was paid based on packets transferred then people would be encouraged to come up with easy/fast transfer of data.

I considered using Bitcoin to pay per packet but I think as the Bitcoin value increases and amount of packets passing through a network increases, you would have to come up with a way to raise and lower prices on your node and have the network automatically find the cheapest route. Maybe even having a setting on your browser to pay extra for increased speed if you want.

That is a big can of worms. You should be familiar with the Network Neutrality debate before proposing Internet pricing schemes.

In my prefered scheme, everybody buys guaranteed bandwidth near the cost of provisioning the wholesale bandwidth. We are talking 64kbps chunks for consumer use: enough to transfer about 16GB in a month if the connection is saturated. The ISP would honor any priority flags on your packets up to that bandwidth. Once you exceed you guaranteed bandwidth, your data would only get through on a best-effort basis (like the current Internet for the most part). You would not be charged per packet or GB untill you exceed your guaranteed bandwidth allocation: other users are able to use your provisioned bandwidth when you aren't transmitting anything. Put another way, for every 64kpbs channel, 16GB of data transfer is already paid for. On top of all of that, you would pay a connection fee to maintain a relatively high speed, burstable connection. The connection cost should be about the same, regardless of the actual burstable speed.

The reason my sheme is so complex is that I wanted to derive a fair way for users to get priority traffic. In my scheme, if your only priority traffic is VOIP, you may buy only 1 or 2 64kps channels. If you are hosting a mulitplayer game server with voice chat, you may want 24 channels (1.5Mbps) or more. By paying for the bandwidth you actually use, you also give the ISP an incentive to actually upgrade their infrastructure instead of complaining that you are using it too much. Another important point: unlike with traditional bandwidth caps, my scheme is not punative. In my fantasy world, overages would be charged on the order of pennies per GB rather than dollars per GB.