Distributed Trust Computations for Decentralized Systems

University Of Maryland, College Park, College Park MD

Investigators

Abstract

Abstract This project will investigate the problem of trust computation (or trust inference) within distributed and completely decentralized systems. Work will specifically focus on trust inference in peer-to-peer networks for distributed resource sharing, where no central authority is assumed. Most systems of this sort are designed based on the assumption that a large fraction of users are "altrusitic" while a smaller percentage of users may be arbitrarily "malicious". This research is intended to address the more realistic case where all users are "selfish" (or "rational"), and are motivated by their own self-interest. This work aims to develop appropriate models for analyzing trust inference in distributed systems; to design efficient protocols and mechanisms to ensure cooperative behavior by the users of the system; and to implement such protocols (and test their effectiveness) within the NICE framework developed at the University of Maryland as well as other existing peer-to-peer applications. Graduate and undergraduate students will be involved in all aspects of this research. Software developed as part of this research will be freely distributed for public use.