Responsive Pricing in the Internet
As the Internet grows, changes are going to be necessary in the control mechanisms that allocate resources. The authors suggest the need to a feedback mechanism to determine demand for internet service at any given time and to price service to the user in response to the demand at any given moment.
In theory, such a system would be a fair allocation of scarce resources, and enable users to tailor their behavior to reflect their own price-sensitivity, creating discrimination in the level of demand that is more accurately linked to the availability of resources.
Definitions of Efficiency
There are two different definitions of efficiency, which can be in conflict: network efficiency relates to the utilization of network resources (particularly bandwidth), whereas economic efficiency refers to the value that individuals place on the availability of service when they wish to have it.
To be efficient in the first sense, the network must run at full capacity at all times, such that network resources are not wasted. In this way, suppliers are not paying the cost of equipment that is sitting idle.
To be efficient in the second sense, the network must run at zero capacity at all times. In this way, capacity exists to provide a user with the full capabilities of the network at the instant he needs it.
Feedback and Adaptation
Feedback is a method of improving network efficiency by signaling the users as to the availability of resources. If this availability also impacted price, the user would have a signal of the capacity of the network at any given time, and could adapt their behavior accordingly, moving less essential activities to off-peak hours.
Should a conflict occur, the result of such a system would be an auction among users, such that the user with the greatest need would outbid those whose needs were lesser (evidenced by the point at which they decide that having the connectivity immediately is not worth the price they would have to pay for it), who would presumably return later to perform these tasks when the price of bandwidth was not prohibitive.
Modeling User Adaptation to Feedback
Users will adapt to the availability of network resources based on price - in fact, such is already the case when the "price" for obtaining information is merely the time required to wait for it to download when the network is congested.
If it requires too much time to download a resource, the user will stop trying, and possibly return later to attempt to download it when traffic has decreased. And it is a standing practice for programmers to download large files or run lengthy processes at off-peak hours to take advantage of the greater efficiency of the network in those times.
When it comes to perishable information - information that is needed right away, but will not be of any use at a later time - the user may be compelled to pay a premium price to have the information right away, or forego purchasing it at all (in which case, the user voluntarily decided that it wasn't "worth" the price).
Responsive Pricing Schemes
The author developed a few different models of responsive pricing schemes and ran it against a sample of usage patterns, but I don't know if there's much to be gained from a purely mathematical simulation where a behavioral factor (willingness to pay a higher price for immediate access) is simulated by a logical algorithm that is designed to responds to the very stimulus being tested,.
Objections to Responsive Pricing
The primary rejection to responsive pricing is the (mistaken) assumption that the Internet is essentially "free". This never really was true (it was paid for by tax dollars, even though users were not charged for access).
Others feel that responsive pricing doesn't necessarily mean that resources are allocated to those that need them, but are given to those who can afford them (regardless of their need). The author tries to bat this back, somewhat sloppily - the proper response is that need is subjective, so no matter the allocation, one can always argue that someone who doesn't "need" something is getting it, to the detriment of someone else whose "need" is perceived as being more important.
Another response is that responsive pricing leads to profiteering; charging the customer more than access is objectively worth. This overlooks competition, and the tendency of industries with high profit margins to attract new entrants who will compete on price with established suppliers.
Conclusions
Given that, as the audience of the Internet grows, the availability of bandwidth to serve the needs of the users will become a scarce resource, for which users will compete.
Given that increasing the bandwidth of the backbone is a significant task that cannot be done in immediate response to increases in demand, scarcity cannot be addressed, in the short run, by providing more bandwidth.
The conclusion is that basing the price of access on users' demand will fairly apportion access to those users whose need is greatest, as evidenced by their willingness to pay the most, at any given time, and that this same solution will guide users whose need is less to modify their behavior to obtain the access they desire at a price they are willing to pay.
EN: It remains an interesting theory, even though the point has largely been made moot, as suppliers have been diligent in providing sufficient bandwidth to meet customer needs for quality - and in periods where network congestion occurred, it was not constant enough to merit investment in a long-term solution, and generally resulted in outages on the supply-side (a Web site was not available) rather than the demand-side (an individual wasn't able to connect at all).