In this work, the basic concepts of two branches of game theory, non-cooperative (including leader-follower, two-person, and n-person) and cooperative games, are reviewed and applied to the study of Internet pricing issues. We emphasize the cooperative game (also called the bargaining problem) approach by illustrating through a simple Quality of Service (QoS) model that the leader-follower game solution does not achieve Pareto optimality and both the Internet Service Providers (ISPs) and the users could be better off by cooperation. We further provide numerical examples for realistic applications and show that the leader-follower game may be "unfair" in some sense for some cases. The analysis is extended to different pricing schemes, such as Paris Metro Pricing (PMP) and Pricing with P...[ Read more ]

In this work, the basic concepts of two branches of game theory, non-cooperative (including leader-follower, two-person, and n-person) and cooperative games, are reviewed and applied to the study of Internet pricing issues. We emphasize the cooperative game (also called the bargaining problem) approach by illustrating through a simple Quality of Service (QoS) model that the leader-follower game solution does not achieve Pareto optimality and both the Internet Service Providers (ISPs) and the users could be better off by cooperation. We further provide numerical examples for realistic applications and show that the leader-follower game may be "unfair" in some sense for some cases. The analysis is extended to different pricing schemes, such as Paris Metro Pricing (PMP) and Pricing with Priority, which leads to the same insights and strengthens the robustness of our model and conclusions. For Pricing with Priority, we also accomplish simulation results since there are no formulae for waiting time distributions. In the end, the two ISP case is studied and we do find a Nash equilibrium point for our example. The practical implication of this work is that some regulations or arbitrations will be helpful for reaching a fairer and more efficient solution. Some discussions are provided as a conclusion.