车联网中基于服务的虚拟网络功能放置算法

Service-based virtual network function placement algorithm in Internet of vehicles

为满足车联网中海量数据的采集、传输以及对这些数据的快速处理的需求,可采用移动边缘计算(MEC)技术。本文考虑移动边缘计算中基站连接方式和物理资源的特点,对边缘服务器的部署问题进行了分析,以部署成本和网络时延为优化目标,划分基站集群,并使用整数线性规划(ILP)建立模型。为了获得运行效率更高的边缘服务器部署方案,本文使用分支定界算法和启发式贪婪算法获得优化模型的近似最优解。实验评估结果显示,分支定界算法和启发式贪婪算法最高可以把边缘服务器部署算法运行时间减少37.6%。此外,本文分析了用户服务器请求数量和用户服务优先级对算法运行时间和边缘服务器运行成本的影响。

In order to meet the needs for the collection and transmission of massive data in the Internet of vehicles and the fast processing of these data,mobile edge computing(MEC)technology can be used.This paper considers the characteristics of base station connection methods and physical resources in mobile edge computing,analyzes the deployment of edge servers,optimizes deployment costs and network delays,divides base station clusters,and builds models using integer linear programming(ILP).In order to obtain a more efficient edge server deployment scheme,this paper uses branch and bound algorithm and heuristic greedy algorithm to obtain the approximate optimal solution of the optimization model.The experimental evaluation results show that the branch and bound algorithm and the heuristic greedy algorithm can reduce the running time of the edge server deployment algorithm by up to 37.6%.Moreover,the paper analyzes the impact of the number of user server requests and user service priority on the algorithm running time and edge server running cost.