Resource Management Technique for IoT in Fog Computing Supported by Distributed SDN

Abstract

Internet of Things (IoT) refers to the interconnection of a very large number of heterogeneous-limited-resource-devices that senses and collects information about their environments. The traditional method to solve the resource scarcity problem in IoT is to leverage the required resources from cloud environment. Things continuously send the requests to the cloud through internet connection. But this is not an optimal solution due to the latency and bandwidth expensive, so the optimal solution is a fog computing. The idea behind the fog computing is moving the resources to the network edge to be close to the IoT devices. In this paper we propose efficient resource management technique based on software defined network (SDN) capabilities to enhancing the QoS of IoT by exploiting the collaboration between the fog and cloud computing. We propose architecture of clusters of fog devices controlled by distributed SDN controllers. In addition the proposed architecture contains central SDN controller connects to all distributed SDN controllers and all cloud servers. So that it contains global view of the network. This paper investigates many issues: task scheduling, mitigate the load, resource discovery and resource selection to reduce the response time and guarantee execution all the hard real time tasks within their deadlines and produces the best effort to execute the soft real time tasks to reduce the penalty.