Implementation of Multiprocessing and Multithreading for End Node Middleware Control on Internet of Things Devices

Abstract

Previously, an educational robot system was built by incorporating Internet of Things (IoT) elements. Over time, this educational robot has been implanted with a middleware. Middleware has a role in receiving command data from the real-time database, access sensors, actuators, and sending feedback. Middleware contains protocols that translate commands between high-level programming and Raspberry Pi hardware. The focus of this research is to improve the performance of the middleware to pursue processing time efficiency. For this reason, it is necessary to implement multiprocessing and multithreading in handling several tasks. The CPU division has been adjusted automatically to not work on just one core or block of memory. Several program functions can run in parallel and reduce program execution time efficiently. The tasks handled are sensor reading and actuator control in the form of a motor. Testing has been carried out to perform multiprocessing and multithreading tasks to process six sensors and five actuators. Multiprocessing requires an average of 1.00% to 15.00% CPU usage and 2.70% memory usage. Meanwhile, multithreading involves an average of 1.00% to 71.00% CPU usage and 3.30% memory usage.