Imperialist competitive based approach for efficient deployment of IoT services in fog computing

Although Quality of Service (QoS) and cost reduction are the main achievement of using resource rich cloud computing in IoT environments, the centralized architecture of cloud computing paradigms and long distance between IoT applications and resources causes to some inefficacy especially in real time applications. Hence, fog computing was joined into cloud computing as a new paradigm to overcome these limitations. Fog computing can provide required resources for IoT devices at the edge of network without involving the cloud. This causes processing, analysis, and storage be closer to the clients and data creation locations, thus efficiency can be improved. Each real time IoT application includes a set of services with different QoS requirements. The resources required for these services can be provided by deploying on fog nodes. This study addresses the IoT Service Placement Problem (SPP) as an autonomous planning model in fog computing. The Imperialist Competitive Algorithm as a metaheuristic approach to solving this problem was developed. Resource distribution is leveraged during allocation process considering fog nodes with sufficient resources because they can host multiple IoT services. The proposed algorithm prioritizes IoT services to reduce delay and solves SPP as a multi-objective problem. Service cost, energy consumption, resource utilization, delay cost and throughput are the specified objectives. In addition, conceptual framework is considered for expressing the proposed autonomous planning model and describing the interactions between the components of the cloud-fog-IoT ecosystem. The proposed algorithm is evaluated by simulation on a synthetic fog environment compared to its counterparts. Experimental results show the proposed algorithm can effectively improve service placement performance 9–17 percent against state-of-the-art algorithms.