In IEEE 23rd International Conference on Data Engineering (ICDE), pages 1400-1404, April 2007.
Acceptance rate: 28%. Number of submissions: 659.
In XML data dissemination networks, data producers publish XML documents and data consumers subscribe to documents of interest by specifying XPath expressions. The problem lies in routing the XML document throughout a network of XML routers to the interested consumers. To solve this problem we show how to extract XPath-like information from document type definitions at the data producer site. This information is first disseminated throughout the network and then evaluated against XPath filter expressions submitted by data consumers signaling interest in receiving documents. To reduce the amount of information required for storing routing state at each node in the network, we introduce covering and merging techniques for XPath expressions. We maintain the subscriptions in a tree data structure to capture the covering relations. Our experimental evaluation shows that these optimizations reduce routing table size by up to 90%, improve routing performance by roughly 85%, and reduce overall network traffic by up to approximately 35%.
Readers who enjoyed the above work, may also like the following: