Neighbourhood maps: decentralized ranking in small-world P2P networks

Reputation in P2P networks is an important tool to encourage cooperation among peers. It is based on ranking of peers according to their past behaviour. In large-scale real-world networks, a global centralized knowledge about all nodes is neither affordable nor practical. For this reason, reputation ranking is often based on local history knowledge available on the evaluating node. This criterion is not optimal, since it ignores useful data about interactions with other peers. In our approach, evaluations of past history create recommendations between nodes, that will be used to form a network called web of trust. Under the assumption that the web of trust has the ubiquitous small-world property, we propose a simple, scalable and decentralized method, called 'neighbourhood maps', which approximates rankings calculated using link-analysis techniques, exploiting the short-distance characteristics of small-world networks. We test our algorithms using data from the OpenPGP web of trust, a real-world network of trust relationships, and by developing a simple simulation of a file-sharing network using an evolutive approach. Our results show that it is sufficient to have maps having size O (root n), where n is the size of the network, in order to have good results. Copyright (c) 2007 John Wiley & Sons, Ltd.