Real-Time Data Analysis using Fog Intelligence

Abstract

Internet is experiencing an unusual growth primarily because of geometric increase in the number of Internet of things (IoT) devices getting connected every day. The increase in the number of communicating devices is causing a stiff surge in overall data getting generated. In a typical cloud computing environment, this huge voluminous data are transmitted directly to the cloud for storage and computation purpose. This is causing serious pressure on the computing environment and network capacity as well. Thus, there is a need to perform the processing of these huge data closer to the source network so that only the relevant or meaningful data flow to the cloud. This resulted in the adoption of a new computing paradigm known as fog computing which aims at storage and computation nearer to the edge of the network thereby helps in overcoming various cloud computing drawbacks. The paper presents the taxonomy of fog computing and an application where different artificial neural network (ANN)-based models are developed and tested on various datasets such as room occupancy dataset, blood sugar percentage, and light and humidity dataset. The functional effectiveness of the proposed fog architecture and its performance in these adopted scenarios are presented through different experimental results.