Abstract
IoT devices deployed in Smart Cities usually have significant resource limitations. For this reason, offload tasks or data to other layers such as fog or cloud is regularly adopted to smooth out this issue. Although data offloading is a well-known aspect of fog computing, the specification of offloading policies is still an open issue due to the lack of clear guidelines. Therefore, we propose OffFog—an approach to guide the definition of data offloading policies in the context of fog computing. In order to evaluate OffFog, we extended the well-known simulator iFogSim and conducted an experimental study based on an urban surveillance system. The results demonstrated the benefits of implementing data offloading based on OffFog recommended policies. Furthermore, we identified the best configuration involving design decisions such as data compression, data criticality, and storage thresholds. The best configuration produced at least 76% improvement in network latency and 5% in the average execution time compared to the iFogSim default strategy. We believe these results represent a significant step towards establishing a systematic decision framework for data offloading policies in the context of fog computing.
Highlights
Fog computing plays an essential role in building a sustainable IoT infrastructure for smart cities [1]
Despite Fog computing brings advantages to the development of IoT applications such as latency improvement, new challenges related to storage, and local processing power emerges in the context of Smart Cities [3, 4]
Data offloading is a well-known aspect of fog computing applications, the specification of offloading policies is still an open issue due to the lack of clear guidelines
Summary
Fog computing plays an essential role in building a sustainable IoT infrastructure for smart cities [1]. Many research efforts have been made on fog computing in smart cities, but there are still several open challenges towards its concrete realization [1, 2]. Despite Fog computing brings advantages to the development of IoT applications such as latency improvement, new challenges related to storage, and local processing power emerges in the context of Smart Cities [3, 4]. IoT devices deployed in Smart Cities usually have significant resource limitations. A usual solution is to transfer the responsibility to entities with more resources, known as offloading. One particular category is the data offloading between the fog and cloud. The need to transfer the data to the cloud emerges in several cases, being very usual when there is the need for long-term storage of data produced on the network border [5]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
