Framework-Based Development and Evaluation of Cost-Based Native XML Query Optimization Techniques

Reflecting on the history of database management systems reveals that cost-based query optimization has been the dominating method for effectively answering complex queries on large documents. Native XML database management systems provide an efficient infrastructure for storing, indexing, and querying large XML documents. Even though such systems can choose from a huge set of structural join operators and value-based join operators as well as various index access operators to efficiently query XML data, the development of powerful native XML query optimizers is just emerging. Furthermore, it is not known how the aforementioned operators behave in complex XQuery evaluation scenarios, which occur frequently in real-word applications. The extensible, rule-based, and cost-based XML query optimization framework proposed in this work, provides a basic testbed for exploring how and whether established techniques of relational cost-based query optimization (e.g., reordering of join operators) can be reused and which new techniques have to be developed to make a significant contribution for accelerating query execution. Using the best practices and an appropriate cost model that will be developed using this framework, it can be turned into a stable cost-based XML query optimizer in the future.