Xpath-aware chunking of XML-documents

Abstract

Dissemination systems are used to route information received from many publishers individually to multiple subscribers. The core of a dissemination system consists of an efficient filtering engine deciding what part of an incoming message goes to which recipient. Within this paper we are proposing a chunking framework of XML documents to speed up the filtering process for a set of registered subscriptions based on XPath expressions. The problem which will be leveraged by the proposed chunking scheme is based on the observation that the execution time of XPath expressions increases with the size of the underlying XML document. The proposed chunking strategy is based on the idea of sharing XPath prefixes among the query set additionally extended by individually selected nodes to be able to handle XPath-filter expressions. Extensive tests showed substantial performance gains.