Building a PDMS Infrastructure for XML Data Sharing with SUNRISE

Semantic support for data representation as well as a flexible machine-readable format have made XML the de facto standard for Internet applications semantic interoperability. Its applicability is primarily evident in realities where actors are heterogeneous data sources which interact each other for data sharing purposes. This is exactly the scenario envisioned by Peer Data Management Systems (PDMSs), where autonomous sources (peers) model their local data according to a schema, and are connected in a peer-to-peer network by means of pairwise semantic mappings between the peers' own schemas. One of the main challenges in such a semantically heterogeneous environment is concerned with query processing when dealing with the inherent semantic approximations occurring in the data. In this paper we present an instantiation of SUNRISE (System for Unified Network Routing, Indexing and Semantic Exploration) for XML data sources. SUNRISE is a complete PDMS infrastructure which extends each peer with functionalities for capturing the semantic approximation originating from schema heterogeneity and exploiting it for a semantically driven network organization and query routing.