Recent advances in sensor technologies and expansion of wired and wireless communication protocols enable us to continuously collect information about the physical world, resulting in a rich set of novel services. The ability to infer relevant patterns from these event streams in real-time and at various levels of abstractions to make near instantaneous decisions is crucial for a wide range of mission critical applications ranging from real-time crisis management to security. This project designs, implements, and evaluates a novel complex event processing methodology, henceforth called Complex Event Analytics (CEA). CEA integrates the capabilities of pattern matching from complex event processing with the power of multi-level analysis from static OLAP engines to provide multi-dimensional sequential pattern analysis over high-speed event streams. The CEA Model combines CEP and OLAP techniques for efficient multi-dimensional event pattern analysis at different abstraction levels. Based on interrelationships in both concept and pattern refinement among queries, sequence queries are composed into an integrated event pattern hierarchy. OLAP like operations enable analysts to navigate from one E-cuboid to another in this event analytics space. CEA optimization strategies, including rewriting rules, physical operators, and cost-based search algorithms, achieve scalable event processing. CEA offers high-performance analytics by maximal shared processing of event pattern queries. Experimental studies compare the CEA solution to the state-of-the-art, including traditional stream query systems and customized event engines. Intellectual merit lies in the design, development and evaluation of a novel Complex Event Analytics technology for real-time event stream analysis, -- a perfect middle ground offering both the sophisticated power of pattern matching found in modern event processing systems and the capability of online analytic techniques at multiple levels of abstraction of OLAP engines. CEA impacts society by facilitating a broad range of stream-centric applications ranging from monitoring of hygiene compliance to prevent the spread of infectuous diseases in medical settings to business intelligence processing, and by integrating project activities with education.
For further information see the project web site at the URL: http://davis.wpi.edu/dsrg/PROJECTS/CEA
