Mobility Aware-Task Scheduling and Virtual Fog for Offloading in IoT-Fog-Cloud Environment

The Internet of Things (IoT) is an emerging technology that provides services to any smart device at any place. Due to the large volume of data and dynamicity, the IoT environment faces challenges in terms of overloading and energy consumption. To resolve those issues, we proposed MISSION (MobIlity taSk Scheduling OffloadiNg) method to solve these problems, and for that, we present four phases. (1). History aware handover process, in this stage we manage the mobility of IoT devices to reduce the retransmission rate. The handover is managed by 5G gateway for that we proposed Mobility Aware Proximal Policy Optimization (MAPPO) algorithm with RSS, direction, and distance parameters. (2). Multi criteria based task classification and scheduling, in this stage first the tasks are classified into four types such as real time task, non-real time task, delay sensitive task and resource intensive task and the classified tasks are given to the input for scheduling, for scheduling we consider a priority, maximum response time, size of data, task completion time and energy. Both the classification and scheduling are done by Di-Process Modular Neural Network (Di-MNN). (3). Energy aware task allocation, in this process, first we calculate the weight value of the task using First Fitness based Animal Migration Optimization (FFAMO) which considers processing time, processing cost, throughput, and energy consumption parameters. And then the weighted task is assigned to the optimal fog by using Capacity based Hungarian Assignment algorithm (CH2A) by considering the load of task, waiting time, energy consumption, and distance which allocates the task optimally. (4). Efficient task offloading on virtual fog nodes, during scheduling or task allocation the fog node gets overloaded that time virtual fog is created by the central fog node using graph entropy. Finally, simulation is done by iFogsim which evaluates the performance in terms of completion time, energy consumption, delay, response time, number of unnecessary handovers, offloading time, bandwidth utilization, and fog load and throughput.