Search Results

With advancements in technology over the last ten years, data management issues have evolved from a stored persistent form to also include streaming data generated from sensors and other software monitoring tools. Furthermore, distributed, event-based systems are becoming more prevalent, with a need to develop applications that can dynamically respond to information extracted from data streams. This research is investigating the integration of stream processing and event processing techniques, with expressive filtering capabilities that include queries over persistent databases to provide application context to the filtering process. Distributed Event Processing Agents (DEPAs) continuously filter events from multiple data streams of different formats that provide XML views. Composite events for data streams are expressed using the Composite Event Detection Language (CEDL) and mapped to Composite XQuery (CXQ) for implementation. CXQ is a language that extends XQuery with features from CEDL, including operators for expressing sequence, disjunction, conjunction, repetition, aggregation, and time windows for events. Continuous queries and composite event filters are integrated with techniques for materialized view maintenance and incremental evaluation in condition monitoring to provide efficient ways of enhancing stream filters with database queries. The filtering and event detection load is distributed among multiple DEPAs, with CXQ expressions decomposed to allocate subcomponents of the expression to DEPAs that efficiently communicate in the global detection of composite events. A unique aspect of our research is that it extends XQuery with temporal, composite event features to combine techniques for continuous queries in stream processing, incremental evaluation in condition monitoring, and detection and filtering of composite events, creating an expressive environment for the extraction of meaningful events from multiple data streams with XML views.