Incentive pricing in multiclass systems

In modern communication networks which offer multiple classes of services, an appropriate assignment of service classes to users (or applications) will have a key influence on the performance profile. Differentiated pricing is an important tool for guiding the user's choice. We consider a basic model for a multiclass system that offers multiclass services to multiple types of traffic, and propose a pricing framework which is based on the concept of nominal traffic assignment. Users (or their associated traffic) are categorized into a finite number of traffic types, which are distinct in their performance utilities at the different service classes. The system administrator specifies a required traffic assignment, which associates with each traffic type a nominal service class. Users, on the other hand, choose service classes so as to optimize their own performance. Optimal prices should provide incentives for the users to assign each traffic type to its nominal service class. Our goal is to implement a simple pricing scheme that provides such incentives. We establish necessary and sufficient conditions for the existence of optimal prices and provide an algorithm for their computation. We indicate that optimal prices can tolerate fluctuations in the various parameters. Next, we propose a distributed algorithm, which can be used by the system to compute optimal prices even when it does not have sufficient knowledge of traffic characteristics. We then generalize our analysis to an extended model, which explicitly includes congestion effects.