Abstract:
Internet of Things (IoT) services have gained unprecedented recognition in the past few years
and become critical to support real-time high data rate applications. This will introduce new
challenges to supporting application in terms of latency, bandwidth, and sustained
connectivity, etc. Due to these constraints, it is challenging for the existing centralized cloudcomputing model to supports real-time applications. To cater these challenges, it is necessary
that the data is processed near the end devices. Therefore, in this research, we present a new
Fog computing based 4-tier architecture for the efficient data processing generated by the IoT
devices specifically targeting real-time applications. Under this architecture, the desired data
coming from the cloud is cached at fog nodes and the data generated by the IoT nodes is
processed near the edge devices. We erected a predictive model for fog nodes selection focus
on minimizing the latency and hit-miss ratio by the use of machine learning. Our architecture
achieved better performance in terms of throughput, delay and hit-miss ratios.
