Peer-to-Peer systems have emerged as a cost-effective means of sharing data and services and are offering fault-tolerance and self-adaptation in large-scale environments. However, the efficient location of data objects or services in a fully decentralized, self-organizing, unstructured overlay network remains a challenging problem. Most of the current solutions rely on maintaining global knowledge or generate large amounts of traffic and as a result do not scale well.

In this work we propose adaptive data dissemination and content-driven routing algorithms for intelligently routing search queries in large-scale, unstructured systems. In our mechanism nodes build and maintain content synopses of their objects and adaptively propagate them to the most appropriate peers. Based on the content synopses, a routing mechanism is being built to forward the queries to those nodes that have a high probability of providing the desired results. Through extensive simulation studies of networks of thousands of nodes and for different content synopses propagation strategies, we show that our approach is highly scalable and significantly improves resources usage by saving both bandwidth and processing power.

cd /home