Query Processing for High-Volume XML Message Brokering

XML filtering solutions developed to date have focused on the matching of documents to large numbers of queries but have not addressed the customization of output needed for emerging distributed information infrastructures. Support for such customization can significantly increase the complexity of the filtering process. This chapter develops shared processing to support the customization of output in the context of high-capacity XML message brokering. It compares three different ways of exploiting a shared path-matching engine for this purpose. The results show that the most aggressive of the three in terms of path sharing performs best, when combined with optimizations based on the queries and Document Type Definitions (DTD). This chapter shows how to leverage an efficient, shared path-matching engine to extract the specific XML elements needed to generate customized output in an XML Message Broker. It also presents techniques to optimize the post-processing of the path-matching engine output and enables the sharing of such processing across queries. This chapter evaluates these techniques with a performance study of the implementation.