Research On Window-based Query Processing On XML Data Streams

As a universal markup language, XML has been the de factor standard for the data organization, data presentation and data exchange on the WWW due to its semi-structured characteristics, good scalability and self-describility. With the rich deployment of applications such as Web Service and personal content delivery, more and more information or data has been encoded in XML format and exchanged on the internet, therefor bring up the computation model based on data streams. The combination of the two bring two basic issues-best selection and joining among different data sources, resulting in an urgent need to developing new and efficient techniques to processing queries on XML streams.This dissertation focuses on the query processing techniques on the XML data streams on a thorough analysis of the research status in this field with the purpose of enriching the XML querying architecture and solving the problems mentioned above efficiently. New processing and optimization techniques to Top-k and Join queries are presented and the effectiveness of the proposed method is verified by experiments and theoretical analysis. The main contribution of the paper can be summarized in the following points:(1) It presents a method for window-based query processing on XML data stream by extending the XPath with window function with a simple syntax. Compared with the small amount of research work, a better balance is archived between the two aspects of the expressive power of the query language and feasibility.(2) I It presents a method to processing Top-k query on XML data stream which integrated the similarity calculation of the XML data with the query in the single pass data scanning process. By using the stack and the priority queue technology, a high processing speed is ensured and feedback speed is improved through results generated progressively and outputted as soon as possible.(3) It presents a new Top-k query processing model to dealing with the potential unboundness issues which is based on the sliding window computation model in which k most relevant results are returned within a window w, where k and w are parameters defined in each query. Theoretical analysis show that by accepting typically very small probability of error, the result queue length is reasonably small and does not significantly depend on stream rates.(4) It presents a method to processing and optimizing join query among XML streams which works in a pipeline and router like way and produces results progressively and gain efficiency under different environmental impact factors.In summary, new techniques and methods different from the existing ones are presented to processing queries on XML streams through an in-depth research and validated by the experiments.