There's a dangerous trend afoot: U.S. wired broadband providers are looking at setting up tiers of service, not based on bandwidth (speed), which many already have, but rather based on amount of data received or transmitted in a given time period. Some Internet service providers are running experiments right now; see Comcast targets bandwidth hogs in test.
It's not, on the face of things, an unfair concept. In some parts of the world, pay-by-the-bit bandwidth (or penalties for overage) is the norm. But adding capped bandwidth services to existing U.S. plans could be disruptive to the economics of the Web here, in more ways than one.
Let's start by looking at an analogous situation in a different industry. Airlines are reacting to high fuel prices by finding ways to tack on charges to tickets. The most recent change: Fees for checked luggage. It doesn't take a genius to predict what some of the consequences of this fee will be on airlines that adopt it: Longer check-in times, since they now become cash transactions; even more crowded overhead bins; and more gate-checked luggage (which is exempt from the fee) cramming the boarding ramp. The new fee will change how we travel and inconvenience everyone, even the people who never check luggage.
A growth in capped broadband services will have similar far-reaching inconveniences.
First, before any ISP starts to charge for usage or shape traffic based on it, they owe it to their customers to provide them with a meter, and preferably historical data, so they know which plan they should subscribe to. This meter really needs to provide users with data showing which applications and sites are their big bandwidth users. Otherwise, customers will have to choose plans based on best guess. But assuming the ISPs implement such a plan (and I'm not sure they will), there will be confusion from many customers trying to decipher the data, and worse, potential domestic discord caused when family members start to quiz each other about their browsing habits.
My biggest concern with penalizing heavy users for bandwidth is that it changes the economic equation not just for users, but for Web site managers and publishers. If every bit has a cost (or even a potential cost), then advertisements may become an unacceptable way to pay for viewing content and using Web sites. The ad blocker software business will flourish, and the economics for content sites and advertising providers could be affected. Hence Google's announcement of a broadband network monitoring service.
Another potential change capped broadband could inspire: Companies that send lots of data into the system and that want to make sure users continue to use it, rate cap or no, could volunteer to foot the extra bill for carriage, exempting their services from users' bandwidth allotments. For example, suppose Netflix wants to be sure its users don't have to worry that watching streaming HD movies will push them over their limit. Maybe Netflix will make a deal with the ISPs to exempt its traffic from counting towards users' limits. But once a deal like that is struck, there goes Net neutrality.
Finally, let's talk quality of service. Some of the plans won't charge users for excess use, rather they'll throttle transmission speed temporarily when high use is detected. That punitive action may affect VoIP (voice over Internet Protocol), games, and other low bit rate, but latency-sensitive applications. Who will field the support calls or foot the bill for lost revenue while people try to figure out why their once-good services no longer work acceptably well?
As I said, overage fees are not inherently unfair. But when you build a business without them and run it that way for years, adding extra fees afterward can be extremely disruptive. And not just to the heavy users of the services, but possibly to the fundamental economics the business relies on.