Edge Orchestration Based Computation Peer Offloading in MEC-Enabled Networks: A Fuzzy Logic Approach

Abstract

Multi-Access Edge Computing (MEC) is a promising candidate to handle the enormous computation demands of many emerging applications and the ever-growing user's quality-of-service (QoS) requirements. However, due to the limitation of computing resource capacity of a distinct edge server, most of the previous studies have proposed a collaboration approach. For collaboration, they considered vertical offloading between mobile with edge computing or edge with cloud computing for taking the advantages of both these technologies. Therefore, these approaches ignored the neighboring edge server having spare computing resources in the same tier. This paper thus proposes edge orchestration based computation peer offloading (EOPO) scheme among the edge servers in the same tier. The main objective is to share the computation resources among the edge servers. Our proposed approach selects the optimal computational node for task offloading based on fuzzy rules. Simulation results corroborate that fuzzy decision based computation peer offloading scheme significantly improves the performance of edge computing. Our proposed EOPO scheme outperformed the two reference schemes which can reduce the average task completion time and task failure rate at approximately 36% and 80.5% respectively when compared with the local edge offloading (LEO) scheme; and at approximately 25.4% and 67.2% respectively when compared with two-tier based offloading between edge with cloud (TTO) scheme.