Identifier-Based Discovery in Large-Scale Networks An Economic Perspective

The design of any network mechanism that requires collaboration among selfish agents could only benefit from accounting for the complex social and economic interactions and incentives of the agents using the design. This chapter presents a broad treatment of the main economic issues that arise in the context of identifier-based discovery on large scale networks, particularly on the Internet. An “identified” object (such as a node or service), referred to as a player, demands to be discoverable by the rest of the network on its “identifier”. A discovery scheme provides such a service to the players and incurs a cost for doing so. Providing such a service while accounting for the cost and making sure that the incentives of the players are aligned is the general economic problem that we address in this work. After introducing the identifier-based discovery problem, we present a taxonomy of discovery schemes and proposals based on their business model and we pose several questions that are becoming increasingly important as we proceed to design the inter-network of the future. An incentive model for distributed discovery in the context of the Border Gateway Protocol (BGP) and path-vector protocols in general is then presented. We model BGP route distribution and computation using a game in which a BGP speaker advertises its prefix to its direct neighbors promising them a reward for further distributing the route deeper into the network. The neighbors do the same thing with their direct neighbors, and so on. The result of the this cascaded route distribution is a globally advertised prefix and hence discoverability. We present initial results on the existence of equilibria in the game and we motivate our ongoing work.