Vague queries on peer-to-peer XML databases

We propose a system, named VXPeer, for querying peer-to-peer XML databases. VXPeer ensures high autonomy to participating peers as it does not rely on a global schema or semantic mappings between local schemas. The basic intuition is that of "vaguely" evaluating queries, i.e., computing partial answers that satisfy new queries obtained by transformation of the original ones, then combining these answers, possibly on the basis of limited knowledge about the local schemas used by peers (e.g., key constraints). A specific query language, named VXPeerQL, allows the user to declare constraints on the query transformations applicable. The system retrieves partial answers, using an intelligent routing strategy, then attempts at combining those referring to the same real-world object.