NickCoons - 5:49 pm on Jun 8, 2006 (gmt 0)

You don't pay for bandwidth for the total network from server to consumer, but for the bandwidth from the webserver to the start of the global internet.

There isn't one entity called "The Global Internet" that all ISPs connect to. The entire internet exists by virtue of all of the independent connections. The ISP that connects my server to the internet connects to various other backbones, and so on, which ultimately go to the end-user requesting data from my site.

If User A requests data from my site, the traffic takes a particular path from them to my server. If User B requests data, that traffic takes a different path entirely, except up to the point where it gets to my ISP. There is no single entity that all traffic passes through that can throttle the bandwidth of those accessing my servers, except my ISP.

The global net is provided on a shared use base and the operators of this part of the net and the IAPs providing the connection between the consumer and the global net are free to throttle your traffic independent of how many you pay your hosting company for the first 500 meters of infrastructure.

But traffic from my server goes through my ISP, then from there might go through AT&T, Savvis, Time Warner, or whoever else to get to the end-user.

My ISP, as with many others, have mutliple connections to the internet. They don't rely on a single provider, as that translates into a single point-of-failure if their one and only upstream provider were to have a problem.

So under this new provision, who would I be expected to pay for faster service? AT&T, Time Warner? Oh wait! I already pay my ISP $X/month for a certain level of service, and they already pay their various upstreams $Y/month for a certain level of service. In other words, I am already paying for a specific amount of bandwidth, and so is my provider, and so is their provider, etc.

So again, how is this different than what already exists and has existed for a long time?