Dynamic service migration in ultra-dense multi-access edge computing network for high-mobility scenarios

Haowei Lin,Xinheng Wang,Juan Zhao,Xiaolong Xu

doi:10.1186/s13638-020-01805-2

Abstract

The multi-access edge computing (MEC) has higher computing power and lower latency than user equipment and remote cloud computing, enabling the continuing emergence of new types of services and mobile application. However, the movement of users could induce service migration or interruption in the MEC network. Especially for highly mobile users, they accelerate the frequency of services’ migration and handover, impacting on the stability of the total MEC network. In this paper, we propose a hierarchical multi-access edge computing architecture, setting up the infrastructure for dynamic service migration in the ultra-dense MEC networks. Moreover, we propose a new mechanism for users with high mobility in the ultra-dense MEC network, efficiently arranging service migrations for users with high-mobility and ordinary users together. Then, we propose an algorithm for evaluating migrated services to contribute to choose the suitable MEC servers for migrated services. The results show that the proposed mechanism can efficiently arrange service migrations and more quickly restore the services even in the blockage. On the other hand, the proposed algorithm is able to make a supplement to the existing algorithms for selecting MEC servers because it can better reflect the capability of migrated services.

Highlights

With the advent of various mobile and Internet of Things (IoT) devices, new types of services and mobile applications are emerging that utilize machine learning (ML) and augmented reality (AR) technology [1]
We propose an algorithm for evaluating migrated services, called Migration Effect Evaluation with quality of service (QoS)-aware (MEEQ), to make a supplement to the existing algorithms for selecting multi-access edge computing (MEC) servers, which is conducive to choose the suitable MEC servers for migrated services (Section 4)
We propose a new mechanism for users with high mobility in the ultra-dense MEC network, called Chain Management with Valuation Adjustment Mechanism (CMVAM), which is extended by Follow-Me Chain (FMC) [28] and MEC architecture proposed in Section 3, efficiently arranging service migrations for users with high-mobility and ordinary users together

Summary

Introduction

With the advent of various mobile and Internet of Things (IoT) devices, new types of services and mobile applications are emerging that utilize machine learning (ML) and augmented reality (AR) technology [1]. They lacked further study on MEC server switching in the high-density environment to enable users with high-mobility and ordinary users to migrate services together without sufficient resources. Peng et al [36] considered the edge user allocation problem as an on-line decision-making and developed a mobility-aware and migration-enabled approach, called MobMig, for allocating users at real-time Both schemes established the learning-based algorithms to choose the suitable MEC servers for the service to be migrated, but lacked further study on QoS of migrated services, leading to fail to choose the most suitable MEC servers. We have added the valuation adjustment mechanism (VAM) application to the MEC servers This MEC networks can still provide stable services and dynamic service migration for common users and high-mobility users together in the ultra-dense network.

Migration effect evaluation with QoS-aware

Results and discussion

Conclusions