Dynamic IoT-Fog Task Allocation using Many-to-One Shortest Path Algorithm

In this paper, we consider a task allocation problem in an IoT-fog environment. We need to allocate a task from the IoT device to any fog node that meets the task's need. A task allocation method must be suitable for the dynamic nature of the network and the fog node condition. Thus, this paper presents an online algorithm and uses the Software-Defined Network. It can serially handle task allocation and get a global network and fog node statistics. The network has many fog nodes with many possible paths, it is important to select an optimal combination of them. So, we formulate the allocation decision as a multi-source single-target shortest path problem on a weighted graph. We propose a joint fog node-link utilization cost and a modified A* algorithm called the Many-to-One Shortest Path algorithm. Our method can give a fast and optimal result. It shows by the experiment result which the proposed method outperforms the previous works.