Game theoretical analysis of incentives for large-scale, fully decentralized collaboration networks

One of the key challenges in peer-to-peer networks is the design of incentives to encourage peers to share their resources. Incentives are necessary in order to exploit the full potential of these systems. The tit-for-tat incentive scheme, as used in BitTorrent for example, has proven to be a successful approach in P2P file sharing systems, where peers have direct relations and share the same kind of resources. However, in P2P systems where different kind of resources are shared between peers with non-direct relations, the design of incentives remains a challenge. In this paper, a large-scale, fully decentralized P2P collaboration network is shown, where peers share not only bandwidth and storage space, but also contribute by editing articles and voting for or against changes. A new incentive scheme is proposed which supports non-direct relations and provides incentives for sharing different kind of resources. The incentive scheme is based on a reputation system that assigns a reputation value to every peer reflecting its previous behavior in the network. Based on this value, the service level is differentiated, i.e. the higher a peer's reputation the better the quality of service it can get from the network. The service differentiation has been analyzed and simulated with rational, irrational and altruistic peers based on game theory concepts.