Cooperative inter-domain distributed resource allocation using stackelberg game

Multiprovider inter-domain resource allocation can be modeled using game theory, since the ISPs act as selfish agents always trying to maximize their revenues. In this paper, we present a new distributive resource allocation framework based on the concept of Slackelberg game theory and dual decomposition. We show that the hierarchy structure game jointly achieve the global optimization through distributed method without any sensitive internal information revealed (such as network topology or self utilities). Under the scheme, we suggest a scalable algorithm for ISPs to implement its policy in a distributed way. And more importantly, our approach is provably Fair and Pareto-efficient. Our scheme and application of game theory between ISPs provides a new, promising direction in distributed algorithmic design, which has heretofore been focused mainly on multicast and ad-hoc network.