Automatic streaming processing of XSLT transformations based on tree transducers

Abstract

Streaming processing of XML transformations is practically needed especially if we have large XML documents or XML data streams as the transformation input. In this paper, we present the design of an automatic streaming processor of transformations specified in XSLT language. Unlike other similar systems, our processor guarantees bounds on the resource usage for the processing of a particular type of transformation. This feature is achieved by employing tree transducers as the underlying formal base. The processor includes a set of streaming algorithms, each of them is associated with a tree transducer with specific resource usage (memory, number of passes), and thus captures different transformation subclass. The input XSLT stylesheet is analyzed in order to identify the transformation subclass to which it belongs. Then the lowest resource-consuming streaming algorithm capturing this subclass is applied. © 2008 Springer-Verlag Berlin Heidelberg.