the price of bandwidth

This article in the NY Times caused quite some uproar in the MMO blogosphere, and it's as good as any other topic to start my blogging career, so welcome ;)
Keen from Keen and Graev's gaming blog got rather enraged about the fact that some major US ISPs plan to develop pricing models which are based on the amount of bandwidth a customers uses per month whereas Tobold seems to think it's a perfectly fine idea. Both posts resulted in interesting discussions, which however mainly displayed the customer perspective. What I'd like to focus on is the marketing perspective, and try to give some insight into whether companies are actually able to implement such pricing models and how the market would react if they did. The underlying marketing question is whether or not it is possible to successfully discriminate prices on the bandwidth market.
Adressing moralistic aspects like "everyone should pay for what they use" would make this post even more lengthy then it already is, so I'm leaving those aside for now.

The fact alone that these pricing models are openly discussed does not necessarily mean that they will work out in the long run. First of all, some customers would definitely be dissatisfied with their ISP and would start looking for alternatives. These customers would then attract companies who are able to offer the same product (= amount of bandwidth) for a lower price.
Whether lower prices in that context are possible depends on the marginal costs a company has per additional bandwidth used by a customer. It's possible that those are not zero, but undoubtedly they're very low. And the fact that telecommunication technologies are still being heavily improved, constantly extending possible limitations, means that the marginal costs of bandwidth are actually decreasing over time.
With non-significant marginal costs of bandwidth and competitors overcharging customers through price discrimination, the market for bandwidth seems an interesting candidate for new entrants, which in turn would keep the prices low.
It's also reasonable to argue that these new entrants would in fact use a pricing model without bandwidth cap because these models have a property which is especially attractive for companies new on the market: simplicity. Instead of having to decide which cap would be appropriate for their household, basing that decision on often crude information about their past usage, consumers could just opt for the 'all-in-one' package. Furthermore, the necessity for the provider to explain different pricing models which consumers might actually find confusing disappears.
Of course, all this depends on the barriers to entry on the American telecommunications market, which I am not really familiar with. Price discrimination works well on monopolistic or small oligopolistic markets, where consumers choices are limited. If entry barriers are high, markets will change less quickly, meaning that discriminating prices would work for a while. Still, the cost situation should provide enough incentive for existing providers to lower their prices below their competitors', provided that it is not possible for companies to fix prices.

In conclusion, I doubt that in a country like the US, with a strong emphasis on free markets, a pricing model that charges heavy users of a low-cost good like the internet significantly more would work out in the long run.
I state that with some confidence because I've seen a similar development here in Germany. While sometimes considered less developed than the US in terms of telecommunications infrastructure, the market itself seems ahead of the US one. Years ago, it was dominated by one monopolistic company, Deutsche Telekom. After the dial-up era with costs per minute, when connections becames faster, the first pricing models implemented were in fact similar to those which are discussed in the US now, with a certain amount of bandwidth included in the basic fee and additional costs for exceeding that amount. Non-cap models, called flatrates, used to be significantly more expensive. But those models, together with a deregulation of the market by the government, allowed small, often regionally active companies to enter the market and actually be able to compete. While the Telekom still remains the biggest player on the market nowadays, they no longer offer any capped models, and aside from the occasional exception you only find flatrates for internet usage on the German market. And they're cheap; I only pay 30 € for a 10 Mbit/s flatrate which also include unlimited telephone access to the german net and basic digital cable TV, which made me raise an eyebrow on Tobold's understanding of the price/bandwidth relationship. ;)