5G边缘计算环境下资源效用与隐私保护权衡的服务迁移方法  

A Service Offloading Method for Resource Utilization and Privacy Preservation Trade-offs in 5G-Enabled Edge Computing

作　　者：李邦源 张春辉 常荣 陈君 许小龙 Li Bangyuan;Zhang Chunhui;Chang Rong;Chen Jun;Xu Xiaolong(Yunnan Power Grid Corporation,Yuxi Power Supply Bureau,Yuxi 653100,China;School of Computer and Software,Nanjing University of Information Science and Technology,Nanjing 210044,China)

机构地区：[1]云南电网有限责任公司玉溪供电局,云南玉溪653100 [2]南京信息工程大学计算机与软件学院,江苏南京210044

出　　处：《南京师大学报（自然科学版）》2021年第4期102-110,共9页Journal of Nanjing Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(61702277).

摘　　要：针对当前5G环境下边缘计算场景存在硬件资源浪费以及服务隐私泄露的问题,本文设计了一种基于资源效用和隐私保护权衡的服务迁移方法(A Service Offloading Method,SOM)以实现边缘节点资源的高效利用与用户的隐私保护.该方法首先对边缘节点在服务迁移场景中资源利用和隐私保护的需求进行了分析与建模,并将其定义为一个多目标优化问题.然后利用改进的强度帕累托进化算法(Improving the Strength Pareto Evolutionary Algorithm,SPEA2)生成问题的可行解,采用简单加权法(Simple Additive Weighting,SAW)和多准则决策法(Multi-Criteria Decision-Making Method,MCDM)从多个可行解中挑选出全局最优解.最后对提出的服务迁移方法SOM进行实验对比并证明其有效性.In view of the problems of hardware resource waste and service privacy leakage in the current 5G environment,a service offloading method(SOM)for resource utilization and privacy preservation trade-offs has been designed to achieve efficient utilization of edge nodes and protection of user privacy.Specifically,resource utilization and privacy protection requirements of edge nodes in service offloading scenario have been firstly analyzed and modeled,and then defined as a multi-objective optimization problem.Afterward,the feasible solutions of the problem are generated by utilizing the improving the strength pareto evolutionary algorithm(SPEA2).Furthermore,simple additive weighting(SAW)and multi-criteria decision-making method(MCDM)are deployed to select the global optimal solution from multiple feasible solutions.Finally,the experimental results show the effectiveness of the proposed SOM.

关 键 词：5G 服务迁移 边缘计算 负载均衡 隐私保护 

分 类 号：G301[文化科学]