An Edge Computing-Based Complex Event Processing Technique for Sensor-Based Systems

Abstract

Complex Event Processing (CEP) on sensor-based systems often uses a mobile gateway agent to forward raw sensor data streams to a remote back-end server. Complex events that are triggered by multiple raw events are then detected at the back-end server. This approach relies on a persistent network connection between the back-end server and the mobile device. This thesis proposes an edge computing-based mobile CEP technique in which CEP is performed on the mobile edge device using an embedded CEP engine and the detected complex events are sent to the back-end server for further processing. A proof-of-concept prototype for this system has been built using a Siddhi CEP engine and a WSO2 server. A thorough performance analysis is performed for comparing the proposed system with the back-end server-based system. The proposed system can handle intermittent network disconnections and leads to reduced user cost and energy consumption for the mobile device.