Latency-Aware Task Partitioning and Resource Allocation in Fog Networks

Smart devices generate various latency-sensitive tasks which need to be processed in real-time or near real-time. The resource-constrained smart devices leverage cloud resources for computing needs. However, the remote geographical position of cloud resources increases the latency. Fog computing provides the computational resources near the smart devices, which can meet the latency demand of many Internet of Things (IoT) applications. Fog computing is a distributed computing paradigm replicating the cloud servers near the origin of the task. However, the fog nodes are also resource-constrained. Hence, resource allocation in fog networks is challenging. Here, we propose a Lagrangian-based task offloading strategy for fog-enabled IoT networks, which partitions the tasks between fog nodes and cloud servers. Moreover, we propose the upper and lower bounds on the required cycle/bit and CPU frequency to allocate the fog node's resources. The extensive numerical results show that the proposed strategy outperforms the existing baseline algorithms.