The Internet is a complex conglomeration of competing networks that are owned and operated by
Internet Service Providers (ISPs) who must balance the technological capabilities of their networks with the economic drivers of their business. We present an optimization-based approach to modeling router-level connectivity that focuses on the causal forces at work in network design and aims at identifying the economic and technical drivers responsible for the observed large-scale network behavior. We show how this model can be used in the context of an “attacker-defender” formulation that identifies the worst-case attack of an intelligent adversary on a router-level network, along with the optimal response of the network to a worst-case attack.