A Minimum Cost Real-Time Ubiquitous Computing System Using Edge-Fog-Cloud

With the development of diverse ubiquitous computing applications, tremendous amount of sensor data is generated. This data requires efficient localized processing and storage. A real-time ubiquitous system requires latency-aware processing to satisfy the deadline of the ubiquitous applications. Performing processing near the network, using Edge and Fog devices, meets this need. The cost of the ubiquitous system depends on the pricing of the processing and storage. In this paper, we present an Edge, Fog, and Cloud layers based ubiquitous computing system, which not only deals with the deadline of a given application but also minimizes the cost of the system. We derive expressions to estimate the cost of computing and storage of the ubiquitous computing system and delay of the network. We demonstrate an application of the analysis in the design of a minimum cost ubiquitous computing system. We propose an algorithm to determine the layers for executing the machine learning techniques required to satisfy the deadline of the system and simultaneously minimize the cost of the network.