移动边缘计算场景下针对资源竞争的服务迁移优化方法  

Service migration optimization method for resource competition in mobile edge computing scenarios

作　　者：王海艳[1,2] 张霖 骆健 WANG Haiyan;ZHANG Lin;LUO Jian(School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China;Jiangsu Key Laboratory of Big Data Security&Intelligent Processing,Nanjing 210023,China)

机构地区：[1]南京邮电大学计算机学院,江苏南京210023 [2]江苏省大数据安全与智能处理重点实验室,江苏南京210023

出　　处：《通信学报》2024年第8期37-50,共14页Journal on Communications

基　　金：国家自然科学基金资助项目(No.62272243)。

摘　　要：针对移动边缘计算(MEC)场景中边缘服务器资源受限导致服务迁移存在资源竞争的问题,基于Lyapunov技术和博弈论,提出了一种针对资源竞争的服务迁移优化方法OMRC-LG。考虑到系统迁移成本有限且当用户数量过多时难以进行轨迹预测,将服务迁移问题建模为迁移成本约束下的最优化问题,并利用Lyapunov技术将最优化问题转化为不需要预测用户轨迹的在线问题处理。为了缓解资源竞争,提出了一种基于博弈论的分布式方法求解在线问题,通过共享用户服务迁移决策以获取准确的边缘服务器可用资源,并不断更新迁移决策,实现服务迁移优化。仿真结果表明,OMRC-LG方法在满足迁移成本约束的同时,降低了平均服务时延。To tackle the problem of resource competition among service migrations caused by limited edge server resources in mobile edge computing(MEC)scenarios,a service migration optimization method for resource competition based on Lyapunov and game theory(OMRC-LG)was proposed.Considering the system's limited migration costs and the difficulty of predicting trajectories when the number of users was large,the service migration was modeled as an optimization problem with migration cost constraints and used the Lyapunov technique to transform it into an online problem without user trajectory prediction.To alleviate resource competition among users,a distributed method based on game theory was proposed.By sharing user service migration decisions,the method obtained accurate information on available edge server resources and would continuously update these decisions to optimize service migration.Simulation results show that the OMRC-LG method can reduce the average service delay while satisfying the migration cost constraints.

关 键 词：移动边缘计算 服务迁移 服务时延 迁移成本 资源竞争 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]