Testbed analysis of multi-fog architecture for interoperable Internet of Things

Fog computing is anticipated to address high latency issues in real-time Internet of Things applications. Deployment of multiple fog nodes reduces the execution time by offloading the overloading task to nearby fog nodes. However, provisioning real-time services demands interoperable communication among different vendor-specific heterogeneous devices and sensing data. The objective of the work is to design a multi-fog semantic network architecture to enable interoperable communication and services from the fog nodes. The proposed architecture implements some frequently used functionalities of the semantic model in the fog nodes. A fog node considers latency and energy as decision-making parameters for selecting an offloader fog node for offloading a task. We evaluate the proposed architecture in an IoT testbed scenario. Raspberry Pi 2 Model B devices are placed at the edge for proposed fog computing operations such as localized semantics, storage, and services. Experiments using our implementation show that the proposed interoperable architecture minimizes the network response time by \(37\%\) $37\%$ (Approx.) compared to traditional cloud computing. The proposed model also scales down the execution time of tasks by \(12\%\) $12\%$ and the energy consumption of the fog nodes by \(10\%\) $10\%$ (approx.) compared to the system without offloading approaches.