At 09:40 AM 8/5/96 -0700, Martin Minow wrote:

Arun Mehta <amehta@giasdl01.vsnl.net.in> writes:

...

True, though even better would be simply to charge you a flat rate. If billing is that expensive, why bother?

Tragedy of the Commons.

Flat rate works only if no single user can use more than a tiny fraction of the total bandwidth.

Using a "Tragedy of the Commons" analysis on telecommunications systems isn't very appropriate. Modern telephone systems have a fairly well-defined instantaneous capacity, do not wear out based on usage, unused capacity doesn't 'store up' for later use, nor do sporadic attempts at excessive use have anything more than a very transitory effect. (fast busy signals.) And in addition, a person doesn't profit in an unlimited fashion by attempting to over-use the telephone: Nobody I know would spend 24 hours per day on the phone if it were free, for example. So there's little motivation to over-use the resource. The Internet is even more "friendly" along these lines than telephone systems: The Internet doesn't "fail hard," denying access when usage is high, it merely slows all access to match the need. There are enough differences that I think Internet deserves an entirely new analysis. Don't worry, it will be also be interesting, from a game-theory perspective, but it will be very distinct from a classic "tragedy of the commons" situation. The current question is how to motivate individuals and companies to invest in improvements to the Internet that will benefit everyone. However, I don't think that will be the limiting factor that it may currently appear to be. Due to the nature of the Internet, there is nothing to prevent a company (such as AOL, Compuserve, or other) from building a shadow version of the Internet, through which all of its customer's traffic will pass until it emerges local to its destination. Customers who appreciate this kind of prompter service will be motivated to pay slightly more and will buy Internet access through that company. So the "commons" won't be quite so "common," and product differentiation will allow choice. Jim Bell jimbell@pacifier.com