Optimal multiclass internet pricing with game theoretical approach

Determining the price to be levied by the internet service provider(ISP) for the usage of network resources has significant effects on network parameters like revenue rates, congestion across the network and provisioning for multiple differentiated levels of service. Current static and dynamic methods of internet pricing do not provide both optimum quality of service (QoS) and a fair share of gain for both ISP and Internet user. Basic concepts of game theory can be applied to the economics of internet pricing for getting better solutions. In particular, a simple model for dynamic pricing using cooperative game theory provides a better solution for both ISP and the user, achieving a better QoS and balanced gains. In this paper, the competition between the ISP and the user is modeled and we find a Nash equilibrium point from which the ISP and the user would not move out without cooperation. The pricing policy used depends on a weighted average queue length at each node. This helps in reducing frequent price variations and is in the spirit of the random early detection (RED) mechanism used in TCP/IP networks. Our work draws on in various ways. We obtain considerable performance improvements using our approach over that in. In particular, our approach exhibits a congestion improvement in the range of 26 to 37 percent, while exhibiting the social welfare improvement in the range of 30 to 38 percent over the scheme in.