GENESIS-CBA: an agent-based model of peer evaluation and selection in the internet interdomain network
1 School of Computer Science, Georgia Institute of Technology, Atlanta, USA
2 The Cooperative Association for Internet Data Analysis, University of California, San Diego, USA
Complex Adaptive Systems Modeling 2013, 1:10 doi:10.1186/2194-3206-1-10Published: 24 April 2013
We propose an agent-based model for peer selection in the Internet at the Autonomous System (AS) level. The proposed model, GENESIS-CBA, is based on realistic constraints and provider selection mechanism, with ASes acting in a myopic and decentralized manner to optimize a cost-related fitness function. We introduce a new peering scheme, Cost-Benefit-Analysis, which, unlike existing peering strategies, gives ASes the ability to analyze the impact of each peering link on their economic fitness. Using this analysis, ASes engage in only those peering relations that can have a positive impact on their fitness.
Our proposed model captures the key factors that affect network formation dynamics: highly skewed traffic matrix, policy-based routing, geographic co-location constraints, and the costs of transit/peering agreements. As opposed to analytical game-theoretic models, which focus on proving the existence of equilibria, GENESIS-CBA is a computational model that simulates the network formation process and allows us to actually compute distinct equilibria (i.e., networks) and to also examine the behavior of sample paths that do not converge.
We find that such oscillatory sample paths occur in about 7% of the runs, and they always involve tier-1 ASes. GENESIS-CBA results in many distinct equilibria that are highly sensitive to initial conditions and the order in which ASes (agents) act.
Our results imply that we cannot predict the properties of an individual AS in the Internet. However, certain properties of the global network or of certain classes of ASes are predictable.