ABSTRACT
As the Internet confronts the belated and somewhat awkward transition from research test bed to commercial enterprise, there has been much recent discussion about the role of pricing in computer networks. This chapter demonstrates that usage-based charging and flat pricing are really two ends of a single continuum: the difference between them is not one of fundamental principle, but merely of degree, and hybrids of the two approaches are likely to be used in the future. It uses the phrase “pricing architecture” to refer to those components of the pricing scheme that are independent of the particular local pricing decisions and reflect nonlocal concerns, such as how receivers rather than senders can be charged for usage and how to appropriately charge multicast transmissions. These architectural issues, rather than the detailed calculation of marginal congestion costs, should form the core of the research agenda. The chapter discusses the economic as well as the mechanistic design issues central to computer network pricing.
