Surviving the Data Glut: The Management of Events Streams in Cyberphysical Systems

The widespread deployment of tagging systems and sensors has resulted in huge streams of events that are automatically captured and must be processed by monitoring and reactive applications. Simple observation events are further aggregated and composed and used to derive complex events of higher level of abstraction, often by enriching the observed events with information from external sources. The processing of event streams requires new mechanisms for event filtering, aggregation and composition, event enrichment and complex event processing, event notification, new correctness criteria and enforcement of non-functional properties, and last but not least, a different software engineering methodology that accommodates the design, test, deployment and governance of event-based systems. We have analyzed a variety of event-based applications in many different domains, ranging from logistics and transportation systems to traffic monitoring and information delivery to the cockpit, from environmental monitoring to e-health systems and ambient assisted living in support of an aging population, and from financial applications to massively parallel multiplayer online games. In this talk we will discuss a selection of these applications and identify their critical features. We will discuss some of the approaches that have emerged to cope with the huge streams of events we are facing and some of the underlying technologies. We will discuss areas in which research is needed and will identify specific research topics where the academic community can contribute in solving some of the problems faced by early adopters of event-based technology.