Descriptive Schema Driven XML Storage.
Unlike traditional relational or object-oriented databases, XML databases require no schema defined in advance. To provide the benefits of a schema in such environments, the notion of descriptive schema (that is also called data guide) was introduced). Descriptive shema is a concise and accurate structural summary of an XML database. It serves as dynamic schema, generated from the base. Descriptive schema helps the user to formulate meaningful queries to the documents without predefined schema. Descriptive schema is also used for query optimisation as a basis for query rewriting, query type inference and physical plan construction.
In this paper we go further and use descriptive schema for organising storage system. Our approach consists in grouping nodes of XML documents in blocks according to their position in the descriptive schema. Thus, descriptive schema plays a role of index structure for path queries that allows us to avoid tree traversal and minimize a number of blocks accessed.Full text of the paper in pdf
Technical Report, 2004.