Leveraging Constrained Devices for Custom Code Execution in the Internet of Things

Abstract

With the ever-growing scale of the IoT, transmitting a massive volume of sensor data through the network will be too taxing. However, it will be challenging to include resource-constrained IoT devices as processing nodes in the fog computing hierarchy. To allow the execution of custom code sent by users on these devices, which are too limited for many current tools, we developed a platform called LibMiletusCOISA (LMC). Moreover, we created two models where the user can choose a cost metric (e.g., energy consumption) and then use it to decide whether to execute their code on the cloud or on the device that collected the data. We employed these models to characterize different scenarios and simulate future situations where changes in the technology can impact this decision.