TIME TO DEATH-BASED SCHEDULING FOR INTERNET OF THINGS IN FOG COMPUTING

Abstract

Internet of things (IoT) is a new modern communication model used to connect a large number of things (devices, humans, and animals) through the Internet. Some of these things have the ability to monitor and sense the environment and collect information that helps in enhancing daily life. Unfortunately, they have limited resources and cannot process, analyze, and store all of the collected data. This issue is solved by the support of cloud computing. However, due to challenges faced in cloud computing, which include transmission time and energy consumption, fog computing has been invented to be an efficient computing system for IoT applications. The IoT sends various types of tasks with diverse importance levels to the fog landscape. Some of these tasks are very important and need to be processed within a given time period (i.e., deadline). This issue has been addressed in fog computing using priority-based (deadline-based) scheduling techniques. However, these techniques are not accurate because they do not reschedule tasks based on current deadlines when new tasks arrive. They do not consider the waiting time of jobs in the queue when updating the deadline's value. This article presents a novel scheduling strategy called time to death-based scheduling for IoT tasks in fog computing. Time to death represents the remaining time to finish the deadline. The simulation results illustrated that the suggested algorithm could help in avoiding the disaster of exceeding the deadlines of time-sensitive tasks. Moreover, it produces good results compared to the delay/priority-aware offloading scheme according to meeting the deadline under various simulation scenarios.