Warning: This may actually be operational too. Given Cogent's (and others) recent pursuit of sub $4/mb/s transit (at least as far as all the bottom feeders and gossippers like to talk about) and the relatively flat cost of a "paid" peering fabric (even on IP8) and the fancy non euclidean name dropping formula costs for cross-connects, the thought of revising current packetized switching model into the historical pay phone model presents itself again. Assuming the Clos switch fabric model, if we simply repudiate all self-evident rights of capitalist business people to compete in a fair market, we should again return to the days of government subsidized monopolies and absolved self-responsibility to avoid the perils of anti-randian unfair competition and a technologically advanced race to the bottom. A "coin-op" best-effort pay as you go switching fabric would be in everyone's best interest (read FAIR and anti-bourgeois classism), and a community designed Clos switch would present no more than 30% blockage on non-peak hours. In this model, people who can't succeed at business could go off and talk about Spam all day, and in their free time they could move cross-connects around, kind of like at the early digital non-NSF funded NAPs. The way I figure it, this would provide a living wage for network engineers that spend all day arguing abstract points in email. You'd simply deposit an American Nickel into the Hawala Peering system. Does this actually improve the quality of the network experience? I dunno. Warning: This may actually have more Merit than other recent posts and the hosting organization of this public discussion list. Yours in contemplation, (f)Rick N. Idiot