A Layered Strategy for Reducing Offloading Latency in Fog Computing

Moving applications from the local Internet of Things devices to neighboring infrastructures has been the highlight of the current distributed Computing era. Due to high latency and network bottlenecks, Cloud Computing has become rather problematic for real-time applications. Accordingly, Fog Computing has been presented as the missing piece to provide along with cloud Computing for virtually all applications. Application virtualization through containers has become more and more popular in the context of Fog Computing as it simplifies live migration and, in particular, offloading of time-critical applications or their selected tasks. However, the measurement study of the overall offloading latency in Fog Computing has not been sufficiently investigated. This article proposes a measurement analysis of the offloading latency of time-critical application tasks from the end device, such as a vehicle, to surrounding Fog nodes based on Docker containers, Kubernetes, and checkpointing. Based on our layered offloading framework, several offloading scenarios have been studied. The experimental results show that by preparing the Fog nodes in advance through careful, easy arrangements, considerable time can be saved when offloading the application to the Fog node within the Fog cluster. We further consider Vehicular Fog Computing (VFC) as a popular example of time-critical Fog computing and assess the suitability of various offloading strategies for VFC use cases.