Skip to Main Content
Multiprovider inter-domain routing can be modeled using game theory, since the domains (or the domain operators) act as players trying to maximize their revenues. Each domain decides what intra-domain connection alternatives it advertises for inter-domain traffic and what the key attributes are of these intra-domain connections (allocation cost, latency, bandwidth or confidence level of availability). The advertised parameters affect the routing algorithm, the selected inter-domain path as well as the payoff of all the operators. As the broadband communication networks become optical, our model assumes a two layer network consisting of autonomous domains, where the optical cross-connects have limited wavelength conversion capability. In this paper we study the dependency of the domain income and the advertised cost of connection alternatives, assuming "cheapest-path" routing. We examined basic scenarios of inter-domain routing and determined Nash and Pareto equilibrium for different network load levels, where provider tariffs and end-user satisfaction were taken into account simultaneously. Beside analytical results, we ran simulations which met our expectations.