Pricing in multiple service class computer communications networks

In order to meet the varied quality-of-service requirements of network applications (such as file-transfer, video, and voice), future packet-switched computer networks will offer multiple service classes. Multiclass network service disciplines, if used properly, utilize network resources more efficiently than single-class service disciplines. Whether or not these multiple service classes are used properly depends on the incentives users encounter when using the network. Thus, the issue of user incentives must be considered in the design and implementation of internet protocols. Two important forms of user incentives in internets are network performance and pricing. We study the role of pricing policies in multiple service class networks. We first argue that some form of service-class sensitive pricing is required in order for any multiclass service discipline to have the desired effect. Borrowing heavily from the Nash implementation paradigm in economics, we present an abstract formulation of service disciplines and pricing policies, allowing us to clearly describe the interplay between service disciplines and pricing policies. Effective multiclass service disciplines allow networks to focus resources on performance sensitive applications. Effective pricing policies can distribute the benefits of multiple service classes to all users, rather than having these benefits remain exclusively with the users of performance sensitive applications. We next analyze the effect of a usage-sensitive pricing scheme on total happiness. Economists have long argued that any scarce resource shared by a large group of self-interested users, requires some form of usage-sensitive incentive structure to discourage overuse. We assess utility in two types of network models: reservation-oriented and reservationless. The former model examines charging in the context of admission control where user demand is a function of the price paid for network service. The charge discourages a fraction of the population with lower benefits so that uses with higher benefits experience less delay in obtaining network service, thereby, experiencing increased utility. The latter model holds offered load constant and varies the data pipe size to reflect different levels of network congestion. In this model, the charge encourages users to select the priority most appropriate to their application. (Copies available exclusively from Micrographics Department, Doheny Library, USC, Los Angeles, CA 90089-0182.)