An N-Tier Fog Architecture for Smart Farming

Current smart farming solutions are either local or cloud-based. The first approach limits the types and scale of available data processing services. In the second approach, the physical and virtual distance between sensors/actuators and the cloud has a significant impact on latency, which leads to quality of service degradation. Another major issue of the all-in-cloud approach is related to data privacy, which is not yet completely regulated. In this paper, we define an N-tier fog architecture and service placement scenarios that concern realistic smart farming applications with data privacy constraints. Then, we introduce a service placement policy that leverages the N-tier fog architecture to cope with availability, deadline satisfaction and data privacy objectives. Finally, we demonstrate the viability of the proposed approach by means of a simulation analysis.