Detecting complex events with real time monitoring infrastructure on event-based systems

Nowadays, in many recent applications; the data size become too large to be managed within the limited storage environments. In such applications, the data is a streaming data and has a continuously flowing data format. Queries on streaming data are performed using declarative queries, as opposed to queries made on classical databases. Declarative queries provide the ability to filter on a continuous data stream. In addition, they provide the ability to follow the data that match the conditions in the queries. Systems with a data structure in a continuous data flow format can be considered Event-based Systems, in which unit data can be considered as a real event. In such systems; real-time detection of complex events that will trigger business processes is becoming an important need. Within the scope of this research, a Complex Event Monitoring Infrastructure is proposed to address this deficiency. The proposed infrastructure provides the detection of complex events, which form the preconditions for initiating business processes in Event-based Systems. In order to demonstrate the usability of the proposed infrastructure, a prototype application has been developed for e-commerce web sites, where the click-stream data is generated in a streaming data format. Prototype application continuously monitors customer click-stream data collected from the Web site, where the trade takes place (e-commerce environment). Each click-stream data, labeled with a timestamp, is considered as an event data. By combining the basic events, occurring within a certain time period, complex events are identified and the appropriate business processes for corresponding commercial activities are triggered. The developed application was evaluated from the perspective of performance and scalability, and positive results were obtained indicating the usability of the proposed infrastructure.