An Efficient Task Scheduling Algorithm in the Cloud and Edge Collaborative Environment

With the advent of the 5G era and the accelerated development of edge computing and Internet of Things technologies, the number of tasks to be processed by mobile devices continues to increase. Edge nodes become incapable of facing massive tasks due to their own limited computing capabilities, and thus the cloud and edge collaborative environment is produced. In order to complete as many tasks as possible while meeting the deadline constraints, we consider the task scheduling problem in the cloud-edge and edge-edge collaboration scenarios. As the number of tasks on edge nodes increases, the solution space becomes larger. Considering that each edge node has its own communication range, we design an edge node based clustering algorithm (ENCA), which can reduce the feasible region while dividing the edge node set. We transform the edge nodes inside the cluster into a bipartite graph, and then propose a task scheduling algorithm based on maximum matching (SAMM). Our ENCA and SAMM are used to solve the task scheduling problem. Compared with the other benchmark algorithms, experimental results show that our algorithms increase the number of the tasks which can be completed and meet the latest deadline constraints by 32%-47.2% under high load conditions.