My guess is that a significant cost difference will last for a long time but will become much less dramatic than at present in the medium term. In the short term (and within the planning horizon of anyone thinking about introducing postgresql hosting) I would guess there could even be a relative widening of the gap as regards database oriented web hosting. This is because a lot of small ISPs will be driven out of the dialup access market by large ISPs/telecos and turn to content provision/hosting. They won't have the one way dialup traffic balancing their other way web hosting traffic and won't be able to get bandwidth for web hosting at rates competitive with north america from the remaining ISPs/telecos who will be few enough to be able to price discriminate. eg Telstra (Australia's near monopoly teleco) has just removed their per MB charge for cable modem access and introduced a flat rate for residential users. This is priced well above the mass market but is sufficient to enable significant numbers of people to experience entirely different internet access than from a dialup modem. They still charge a rate per MB for "business" use and design the system to make it more difficult for people to run their own web services over it (no permanent IP address, random disconnects etc). This enables them to sell web space to businesses at extortionate rates. Medium term this could accelerate the mass penetration of the internet in Australia and thus the general collapse of national barriers, including cost differences for services that can run from anywhere. Short term this could result in a lot of people realizing they need cheap web hosting and that they can't get it in Australia.
These are of course just guesses.
