Assessing the assumptions underlying mechanism design for the Interne

Steven Bauer, Peyman Faratin and Robert Beverly
Proceedings of First Workshop on the Economics of Networked Systems (NetEcon),
Ann Arbor, MI, June 2006.

The networking research community increasingly seeks to leverage mechanism design to create incentive mechanisms that align the interests of selsh agents with the interests of a principal designer. To apply mechanism design, a principal designer must adopt a variety of assumptions about the structure of the induced game and the agents that will be participating. (We focus in this paper on assumptions regarding agent preferences and non-repeated vs. repeated games.) As we demonstrate, such assumptions are central to understanding the degree to which theoretical claims based upon mechanism design support architectural design decisions or are useful predictors of real-world system dynamics. This understanding is central to integrating the theoretical results from mechanism design into a larger architectural discussion and engineering analysis required in networking research. We present two case studies that examine how the valid theoretical claims of [7, 18] relate to a larger, architectural discussion. We conclude with a discussion of general criteria for designing and evaluating incentive mechanisms for complex real-world networks like the Internet.

[PDF(83KB)] [BibTeX]

[ Return to publications ]