Constructing Finite State Automata for High-Performance XML Web Services.

Abstract

This paper describes a validating XML parsing method based on deterministic finite state automata (DFA). XML parsing and validation is performed by a schema-specific XML parser that encodes the admissible parsing states as a DFA. This DFA is automatically constructed from the XML schemas of XML messages using a code generator. A twolevel DFA architecture is used to increase efficiency and to reduce the generated code size. The lower-level DFA efficiently parses syntactically well-formed XML messages. The higher-level DFA validates the messages and produces application events associated with transitions in the DFA. Two example case studies are presented and performance results are given to demonstrate that the approach supports the implementation of high-performance Web services.