Cooperating Peers for Content-Oriented XML-Retrieval

Semi-structured documents formatted with the extensible markup language (XML) are gaining wide use by a whole range of applications including E-Commerce, E-Business, EScience, Digital Libraries (DL), File Sharing, and in the last years especially by applications for Peer-to-Peer (P2P) systems. P2P architectures have been identified as an efficient means of ad-hoc collaboration and information sharing among large, diverse, and dynamic sets of user. However, current P2P search engines for XML-documents lack the use of information retrieval methods to efficiently search XML collections for relevant information. This article proposes a search engine for P2P systems that applies an extension of the vector space model and exploits structural information to compute relevance of XMLdocuments, and thus may significantly improve retrieval performance. We concentrate on the cooperation of peers that perform a distributed query execution through cooperated retrieval and ranking of dynamic XML documents. The interaction between the participating peers is based on a structured P2P-network and uses an adaption of the DHT-algorithm Kademlia.