Pricing versus Admission Control in Multi-class Loss Networks

In this paper, we study the relative roles of optimal pricing and call admission control (CAC) in network revenue performance optimization from a network provider´s perspective. We consider a multi-class loss network where each traffic class has a differentiated quality of service (QoS), as characterized by a per connection bandwidth, and an associated static charging price which decides the arrival rate of the service demand. We formulate a framework that views pricing and call admission control (CAC) as two stages of a unified call level resource allocation process, and subsequently study the impact of pricing and CAC both analytically and numerically. We show that both an optimal pricing and an optimal CAC can improve the performance, and the optimal pricing plays a significant role. Moreover, in large networks with many small users, the performance resulting from the greedy CAC and its corresponding optimal pricing approaches the upper bound performance asymptotically, and more complicated joint optimization of pricing and CAC does not provide an additional performance gain.