Contribution to improving the performance of XML complex data warehouses

Abstract

Many models and optimization techniques have been proposed in the literature. We have presented in our research the main techniques to improve the performance of data warehouses, such as indexing, materialized views, and fragmentation with its types. The work presented in this thesis is part of the physical design of the XML complex data warehouse. We have proposed a query optimization approach based on the concepts of Hash Fragmentation, which we called XHaFrag short for "XML Hash Fragmentation", it is based on the fragmentation of the facts document only, because it is the largest document, and the most updated among the rest of the documents (dimensions). This approach is characterized by the simplicity and clarity of its principle, ease of application and implementation. It is also characterized by diversity of function, and non-compliance with one hash function, and the possibility of working with various fields or properties of facts document. This approach was subject to experiment on the benchmark XWeB, which proved its efficiency and its usefulness and the experiment also proved that it is an idea that can be developed and applied.