Sang Ki Kim
The Internet, Quality of Service, Dynamic Pricing, Admission Control, Diffserv Network, Network Simulation
In order to manage the moral hazard problem regarding the congestion situation of the Internet, many researchers have studied dynamic pricing. However, most dynamic pricing studies rely on an over-simplified demand model, especially for emerging multiple QoS class environments. Thus, we generalize a demand model of an Internet pricing scheme for a multi-class QoS network. We also propose dynamic admission control to supplement our pricing model. In our admission control scheme with a movable boundary, we consider social welfare as the main objective of the model and therefore regard the congestion cost of customers as well as the provider’s revenue as the goals to be tackled. A simulation demonstrated the viability of our model, showing that dynamic admission control was superior to a fixed bandwidth allocation scheme through a comparison of the net value between the revenue of new QoS traffic and the increased congestion cost to existing best-effort traffic.
Important Links:
Go Back
