Project Overview

We study a collection of heterogeneous XML databases maintaining similar and related information exchanging data via a peer to peer overlay network.  In a highly dynamic P2P network, it is unrealistic that a peer which enters the network is forced to agree on a global mediated schema, or to perform heavyweight operations to map its schema to the neighbors schemas.  Yet, users/applications wish to query the collection of databases via one peer using its schema.  We have recently developed HePToX (pronounced Hep Talk), a HEterogeneous Peer TO peer Xml database system.  A key idea is that whenever a peer enters the system, it establishes an acquaintance with a small number of peer databases, possibly with different schemas.  The peer administrator provides correspondences between the local schema and the acquaintance schema using an informal but intuitive notation of arrows and boxes.  We developed a novel algorithm that infers a set of precise mapping rules between the schemas from these visual annotations.  We pinned down a semantics of query translation given such mapping rules, and also developed a novel query translation algorithm for a simple but expressive fragment of XQuery, that employs the mapping rules in either direction.  The translation algorithm has been proved to be correct, and the utility and scalability of HePToX has also been tested with a detailed set of experiments on top of the Emulab P2P system.