The Internet consists of thousands of autonomous but interconnected networks of different types and business objectives. These networks form a complex economy in which traffic flow is coupled with financial exchanges, routing interacts with business agreements, while new services and applications, pricing schemes, and network protocols or technologies are all interwined in an elaborate web of rather poorly understood interactions. Most of these interactions are local in nature, without coordinated control or regulation, but they often have global impact affecting the performance and reliability experienced by users, the financial viability of network and service providers, and to some degree the global economy.
The high-level objective of this research is to create a scientific basis for the study of inter-network economics. Specifically, we aim to understand the Internet ecosystem from an economic perspective, capturing the interactions between network business relations, the dynamics and structure of the global inter-network topology, as well as the resulting interdomain traffic flow.
In the past, networking research has focused on interdomain topology and routing, but rarely considered how those two issues interact with economics. In parallel, previous work in network economics has used game theoretic models (such as network formation, congestion or pricing games) focusing on the existence and structure of equilibria, but without considering the dynamics of this economy or the interactions between business agreements, global inter-network topology, and interdomain traffic flow. On the other hand, network operators and strategists are called to frequently make important decisions about pricing, peering and provider selection, and routing strategies; these decisions are poorly understood, however, in terms of their local, and especially in terms of their global effects.
We propose a research agenda that is based on the synergism between two approaches: computational/simulation methods and game theoretic methods. Computational methods can investigate elaborate models that capture several aspects of what happens in practice. Game theoretic formulations and analysis can provide more rigorous and general answers but only for simpler models. The two methods can be highly synergistic. Comparing the results between an elaborate computational model and a simplified game theoretic model can show which aspects of the former are not essential. Further, computational methods can explore game theoretic approaches that are rigorous, in terms of the underlying formulation, but not analytically tractable. In this research, we will apply this synergism in three directions that we view as cornerstones in inter-network economics:
Even though the main objective of this research is to create a science of inter-network economics, we firmly expect that a deeper understanding of this area will have a broad and major impact in network operations, network design, network pricing, and possibly in Internet regulation. We expect that our results will be of interest to network operators and strategists, content providers, technologists and policy makers.
Constantine Dovrolis (Spring'09)