Event Processing

Current, data-driven applications have become more dynamic in nature, with the need to respond to events generated from distributed sources or to react to information extracted from incoming data streams. Event process-ing and stream processing have traditionally developed as two separate ar-eas of research. Event processing has its roots in research with active rule processing (Widom and Ceri, 1996) as well as distributed systems (Muhl et al., 2006), with a focus on composite event specification languages and execution issues for detecting, broadcasting, and consuming streams of events. More recently, data stream processing has developed as a new form of data management, with a focus on the continuous execution of queries over data generated from sensors or other sources that emit streams of data that must be quickly analyzed (Golab and Ozsu, 2003; Arasu et al., 2003). Research on data streams mainly focuses on continuous (and potentially infinite) sequences of data, investigating query processing techniques that can be “localized ” to recently received streaming data using sliding win-dows to handle the temporal aspects of the stream. Continuous queries for streaming data are similar to past work with condition monitoring for per-sistent data (Rosenthal et al., 1989). Condition monitoring has been used in the context of condition-action rules in rule processing environments to de-termine data changes, known as deltas, that affect the truth value of the condition and to incrementally evaluate the query of the condition for effi-ciency