Searching XML Documents by Keywords in Structured P2P Networks

In this paper we present a scheme for searching XML documents by keywords in DHT-based structured P2P networks. Our idea is to construct an inverted index for XML documents in an DHT-based P2P network. To this end, we discuss how to decompose XML documents and store the resulting fragments in a DHT-based P2P network. Given a query in terms of keywords, we have to be able to find respective XML fragments that subsume given keywords. We address this problem by finding SLCAs (smallest lowest common ancestors) for the keywords with the help of the DHT-based inverted index. We also discuss the technique to reduce number of message transmissions among peers by introducing Bloom filters. We show the feasibility of the proposed scheme by experimentations.