Answering XPath queries with search predicates in structured P2P networks.

The new generation of peer-to-peer applications requires advanced semantic search over distributed XML data sources over the internet. However, most current proposals in the literature only support simple XPath query answers in unstructured or structured P2P networks. In this paper, we present an effective and efficient framework to support complex XPath queries with full-text search predicates over schema-less XML data distributed in a OHT-based structured P2P network. We construct two types of routing data synopses, termed structural summary and peer-document synopsis, to route the user query to most relevant peers that own documents that can satisfy the query. To improve the search performance, our lazy query evaluation terminates as early as possible when the top-k query answers are retrieved. We also develop effective strategies to balance the data load distribution in the system. In addition, we introduce novel methods to handle network updates, such as node arrivals, departures, and failures. Finally, we conduct extensive experiments to show the scalability of our system, validate the efficiency and accuracy of our routing data synopses, and demonstrate the effectiveness of our load balancing schemes. [ABSTRACT FROM AUTHOR]