TY - UNPB

T1 - Towards A Lightweight Incentive Scheme for Peer-to-Peer Systems

AU - Kaune, Sebastian

AU - Tyson, Gareth

AU - Liebau, Nicolas

AU - Steinmetz, Ralf

AU - Content NoE (Funder)

PY - 2008/5/4

Y1 - 2008/5/4

N2 - Peer-to-peer (p2p) systems rely on the contributions of peers to operate. In order to not only depend on the altruism of individual peers, a number of incentive systems have been proposed. However many of them are complex and suffer from high overhead. In this work we propose BioTrust, a lightweight incentive scheme. In contrast to other approaches, it does not require reputation histories and can operate with minimal overhead in terms of bandwidth, infrastructure, memory and computation. BioTrust is based on observations of cooperation as observed in biological systems, which show that individuals try to ally themselves with others that can best increase their own standing. Such observations are prevalent in evolutionary systems where individuals must weigh off their own individual needs against the need for cooperation to survive. Using these models, BioTrust has been developed to extract and modify some of the most useful properties of biological systems to work effectively in a p2p environment. It is shown through simulation that this scheme encourages honest peers, which leads to a close to optimal behavior in the system context as well as for single peers.

AB - Peer-to-peer (p2p) systems rely on the contributions of peers to operate. In order to not only depend on the altruism of individual peers, a number of incentive systems have been proposed. However many of them are complex and suffer from high overhead. In this work we propose BioTrust, a lightweight incentive scheme. In contrast to other approaches, it does not require reputation histories and can operate with minimal overhead in terms of bandwidth, infrastructure, memory and computation. BioTrust is based on observations of cooperation as observed in biological systems, which show that individuals try to ally themselves with others that can best increase their own standing. Such observations are prevalent in evolutionary systems where individuals must weigh off their own individual needs against the need for cooperation to survive. Using these models, BioTrust has been developed to extract and modify some of the most useful properties of biological systems to work effectively in a p2p environment. It is shown through simulation that this scheme encourages honest peers, which leads to a close to optimal behavior in the system context as well as for single peers.

KW - cs_eprint_id

KW - 2164 cs_uid

KW - 392

M3 - Working paper

BT - Towards A Lightweight Incentive Scheme for Peer-to-Peer Systems

PB - Technische Universität Darmstadt

ER -