Energy and Time-Effective Computation Offloading for Edge Computing-Enabled IoT Networks

By connecting and integrating diverse devices over a wireless connection, the Internet of Things (IoT) has revolutionized various domains and environments. However, IoT nodes' constrained battery and processing capacity limit their performance, making computation offloading a viable solution. This solution enables the migration of high-demand applications from IoT nodes to the Edge. This process depends on several variables, including the computing power available at IoT nodes, the accessibility of nearby Edge resources, and the connectivity condition between IoT nodes and the Edge. IoT brings several challenges to the application of computation offloading, including the heterogeneity of IoT and the limited resources of its nodes. To this end, we consider energy consumption and execution time measurements of different IoT applications running on physical IoT sensor nodes. Based on these measurements, we propose adaptive schemes that consider the resources available at the IoT nodes, as well as the number and type of available Edge servers to meet the requirements of IoT applications. Finally, we conduct extensive experiments to analyze the performance of the proposed adaptive schemes against other baseline schemes. We also study these adaptive schemes in a network scenario, where we schedule IoT applications based on the network status.