Mitigating information asymmetries to achieve efficient peer-to-peer queries

Querying for a particular data item is perhaps the most important feature to be supported by peer-to-peer network infrastructures, and receives the most research attention in recent literature. Most existing work follows the line of designing decentralized algorithms to maximize the performance of peer-to-peer queries. These algorithms often have specific rules that peer nodes should adhere to (e.g., placement of data items on particular nodes), and thus assume that peers are strictly cooperative. However, in realistic peer-to-peer networks, selfish and greedy peer nodes are the norm, and query strategies degenerate to random or flooding based searches. In this paper, we explore the design space with respect to query efficiency in selfish peer-to-peer networks where nodes have asymmetric information, and apply the signaling mechanism from microeconomics to facilitate the sharing of private information and thus improve search efficiency. We extensively simulate the signaling mechanism in the context of other alternative solutions in selfish networks, and show encouraging results with respect to improving query performance.