Adaptive event tree-based hybrid CEP computational model for Fog computing architecture

With the bloom of data generation devices at the network-edge, obtaining data intelligence in real-time posed problems due to network incompetence. In this paper we present a novel computation model based on Fog computing that take advantage of its low latency computing capabilities. A proof of concept is designed on top of the computation model as a layered system architecture that consists of sensors as data sources and complex event processing as the computation model in a Fog node. An event tree based scheduling system is proposed to mitigate the event dependencies. We also introduce a Fog to Cloud gateway that schedule data to the Fog or to the Cloud using a rule engine and system resource prediction. We evaluate the system against a traditional Cloud based computation model to validate its effectiveness and real-time computation capabilities.