Funding source: NSF CNS-1017064. Period of performance: June 1, 2010 - June 30, 2013.
The Internet consists of thousands of autonomous systems (ASes) that voluntarily form bilateral (sometimes conditional) interconnection agreements to provide end-to-end reachability. These interactions between networks are local, without centralized control or regulation, but they often have global impact on the performance and profitability of network and service providers. Despite much recent interest in the economic aspects of the Internet, such as network interconnection (peering), pricing, performance, and the profitability of various network types, two historical developments contribute to a persistent disconnect between economic models and actual operational practices on the Internet. First, the Internet became too complex - in traffic dynamics, topology, and economics - for currently available analytical tools to allow realistic modeling. Second, the data needed to parameterize more realistic models is simply not available. The problem is fundamental, and familiar: simple models are not valid, and complex models cannot be validated. We propose transformative progress in both dimensions: creating more powerful, empirically parameterized computational tools, and enabling broader validation than previously possible. We will use measurements of interdomain traffic, topology dynamics, routing policies and peering practices as input to ITER, a detailed model of AS interconnection. With ITER, we take an approach that is computational rather than analytical. ITER takes as input the interconnection policies of various network types, interdomain traffic demands, routing policies, geographical constraints, and pricing/cost factors, and computes an equilibrium - a state where no network has the incentive to change its connectivity. We propose a two-pronged approach to validating ITER. First, we will verify that ITER can reproduce known macroscopic properties of the Internet AS topology. Second, we will use historical, publicly available financial and topological data to verify that ITER can reproduce known trends in the evolution of the Internet. We will then use ITER to study various interconnection practices, the stability and dynamics of interdomain links, and economic properties of the resulting equilibrium. Intellectual merit: We propose an approach grounded in empirical measurements of macroscopic Internet topology, traffic demand, routing policies, and peering policies. First, we will study the evolution of transit and settlement-free links of ASes, attempting to infer the economic incentives and policies of ASes that underlie topology dynamics. We will extend our previous work on the evolution of transit links over the last 10 years by studying the evolution of settlement-free peering links. Second, we will use flow-level traffic measurements collected at multiple vantage points to infer properties of the interdomain traffic matrix. Third, we will use a combination of active and passive measurements to study the economic implications of routing policies as well as peering policies used by ASes. The data promise to reveal important, and thus far elusive, insights into the economic implications of topology dynamics, interdomain traffic characteristics, and routing policy, but they will also inform the parameterization of ITER, our proposed model of AS interconnection and dynamics. We will validate ITER using publicly available historical financial data, BGP data, and information about the changing business roles and strategies of various networks over time. As an application of ITER, we will simulate several plausible "what-if" scenarios relating to traffic, peering strategies, geography, and peering/cost structures and explore the effects of these factors on macroscopic stability and performance characteristics. Broader impact: The proposed research will yield not only deeper, empirically grounded interpretation of available data on the most opaque sub-discipline of network research - internetwork economics - but also a broad understanding of how economic forces induce, as well as result from, topology dynamics and architectural evolution. We have structured our computational approach to enable new transformative research in complex network modeling, including broader issues such as pricing and policy. The educational side of the project will integrate Internet economics in two Georgia Tech courses, while a PostDoc and a PhD student will graduate as experts in that emerging and rather sparsely populated area. Further, our data and methods will be publicly available and regularly presented to both the research community as well as operator and policy forums, e.g., NANOG, FCC.