Can't this be built without the servers?

Why not take a normal bittorrent client and have it determine the number of hops between it and the remote systems, then preference taking the nearest ones?

I understand that there are some things that the client would have a very hard time determining (congestion, dis-joint networks that appear local, etc) but this too should give quite a lot to both the consumer and their ISP. The consumer would get faster downloads (assuming people within their same MAN were also seeding the same files), and the ISP would get a lesser ding to their carrier trunk.

I can see why the carriers like this - they can also tell the client software to use different routes at different times of the day, not to mention the ability to monitor what people are looking for and sell *that* data to the highest bidder.

Dan "I'm Not Synical...but close"

Click to read the article this is in response to.