An Agent-based Model for the Evolution of the Internet Ecosystem

We propose an agent-based model for the evolution of the Internet ecosystem. We model networks in the Internet as selfish agents, each of which tries to maximize a certain utility function in a distributed manner. We consider a utility function that represents the monetary profit of a network. Our model accounts for various important constraints such as geography, multihoming, and various strategies for provider and peer selection by different types of networks. We implement this model in a simulator, which is then used to solve the model and determine a "steady-state" of the network. We then present a set of "what-if" questions that can be answered using the proposed model and by studying the properties of the resulting steadystate.