Improving XPath Query Execution in P2P XML Storage by Using a Decentralized Index

1 KOM, TU Darmstadt, Merckstrasse. 25, 64283 Darmstadt, Germany {pussep,liebau,steinmetz}@kom.tu-darmstadt.de 2 Fraunhofer IPSI, Dolivostr. 15, 64293 Darmstadt, Germany [email protected] Abstract. Today, information is managed incresingly in dynamic communities on the Internet. Here, peer-to-peer communities in which users host data by contributing their resources is a very promising alternative for centralized hosting. Managed data is often represented in XML and requires a high-level query language, where XPath is a good candidate. In this paper, we present a decentralized XML index which enables efficient XPath queries on large documents stored in p2p systems. Unlike other approaches, we do not rely on a specific overlay as our solution is able to work on top of any structured overlay which provides common put and get operations. Our approach is combined with P2P XML Storage, which stores arbitrarily large XML files in p2p networks efficiently. Evaluation has proven that our index improves the XPath query performance regarding both the execution time and the number of messages by orders of magnitude.