Recent research on Distributed Event-based Systems (DEBS) has focussed on event correlation, which is the task of processing events to identify meaningful patterns of events in the event cloud. In DEBS, software components communicate by generating, disseminating and receiving event notifications, which reify and describe the event. Several parts of an event notification are context-sensitive, depending on where the software component producing the event is deployed, the communication infrastructure available for event dissemination etc. Event contexts may be added during the production of an event (e.g. by the runtime system executing the component) or during dissemination (by a middleware) and play an integral part in event correlation. Examples of contextual information include physical time, logical time, geographical coordinates, information about the source of events, digital signatures, etc. EventJava , an extension of Java with advanced support for event correlation, explicitly integrates the notion of event context, thereby allowing a programmer to customize the way in which events are ordered, propagated and correlated with other events. In this paper, we explain why contexts are indispensable to DEBS, present an overview of EventJava and illustrate the use of contexts through programming examples.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below