Efficient XML query processing in mediators

Abstract

With the advent of XML as the standard for data representation and exchange over the Internet, there are ever increasing applications; E-commerce application in particular, using XML as a mediator model. It raises the issue of how to evaluate XML queries efficiently in mediator systems. The authors address the problem of multiple query optimization in an XML based mediator. Because the XML query language differs from SQL/OQL in its pattern matching and regular path expression constructs, the method of identifying common subexpressions of multiple XML queries is different accordingly. We present the steps to identify common subexpressions. To obtain the global query plan, we propose some heuristics to reduce the whole search space. By using a statistical method to estimate the usage frequency of the common subexpressions in XML queries, a selective caching and materializing of remote data can be accurately made, and this will greatly improve query response time.