Towards an automatic deployment model of IoT services in Fog computing using an adaptive differential evolution algorithm

Nowadays, fog computing has joined cloud computing as an emerging computing paradigm to provide resources at the edge of the network, as centralized clouds face challenges such as delay to process the unprecedented volume of data generated by Internet of Things (IoT) devices. Fog computing ensures the processing of real-time IoT applications at the edge of the network with low delay, as there is no need to transfer the entire data to a remote cloud. However, the main challenge is to deploy IoT services as components of IoT applications on fog nodes. Fog nodes are heterogeneous, distributed and resource-constrained, and this motivated us to solve the IoT Fog Service Placement (FSP) problem as a multi-objective optimization problem with evolutionary approaches. Here, we develop an Adaptive Differential Evolution (ADE) algorithm to solve FSP that originates from the MAPE-k (Monitor-Analyze-Plan-Execute over a shared knowledge) autonomous model. ADE considers a reproduction policy based on differential evolution-current-to-best, whose parameters are adjusted adaptively. The proposed method, ADE-FSP, transforms the multi-objective problem into a single-objective optimization problem with the perspective of minimizing deadline violation, resource loss and service cost as well as maximizing resource usage. Meanwhile, ADE-FSP ensures the automatic and efficient deployment of IoT services in the fog environment by considering the priority of services and the distribution of resource consumption. We analyze the proposed ADE-FSP from different perspectives on a simulated fog environment. Experimental results show that compared to state-of-the-art algorithms, ADE-FSP significantly improves delay (up to 5.6%) and resource usage (up to 13.2%).