Optimization of Distributed SgmlQL Queries

SgmlQL is a SQL like language to manipulate SGML or XML documents distributed over a network. In this article we present the query optimizer that we developped for this language. We distinguish between local queries, which access documents stored in the site from which the query is issued, and distributed queries, which access documents stored in sites differents from the one from which the query is issued. Optimisaton of local queries is based on a set of algebraic transformations analogous to those designed for relational or object oriented query processing. Optimization of distributed queries consists in splitting a query into subqueries each one depending on a single site where it will be processed. Finally we present some performance measurments.