Bertrand competition in networks.pdfVIP

Bertrand competition in networks Shuchi Chawla1?, Tim Roughgarden2?? 1 Computer Sciences Dept., University of Wisconsin - Madison. shuchi@. 2 Department of Computer Science, Stanford University. tim@. Abstract. We study price-of-anarchy type questions in two-sided mar- kets with combinatorial consumers and limited supply sellers. Sellers own edges in a network and sell bandwidth at fixed prices subject to capacity constraints; consumers buy bandwidth between their sources and sinks so as to maximize their value from sending traffic minus the prices they pay to edges. We characterize the price of anarchy and price of stability in these “network pricing” games with respect to two objectives—the social value (social welfare) of the consumers, and the total profit obtained by all the sellers. In single-source single-sink networks we give tight bounds on these quantities based on the degree of competition, specifically the number of monopolistic edges, in the network. In multiple-source single- sink networks, we show that equilibria perform well only under additional assumptions on the network and demand structure. 1 Introduction The Internet is a unique modern artifact given its sheer size, and the number of its users. Given its (continuing) distributed and ad-hoc evolution, as well as emerging applications, there have been growing concerns about the effective- ness of its current routing protocols in finding good routes and ensuring quality of service. Congestion and QoS based pricing has been suggested as a way of combating the ills of this distributed growth and selfish use of resources (see, e.g., [5, 7, 8, 10, 12]). Unfortunately, the effectiveness of such approaches relies on the cooperation of the multiple entities implementing them, namely the owners of resources on the Internet, or the ISPs. The ISPs’ goals do not necessarily align with the social objectives of efficiency and quality of service; their primary objective is to maximize their own market share and