基于流量感知的网络切片资源智能分配方法  

A Traffic-aware Intelligent Allocation Method of Network Slice Resource

作　　者：郭红峰[1] 黄继宁[1] GUO Hongfeng;HUANG Jining(China Mobile Group Guangdong Co.,Ltd.,Guangzhou 510623,China)

机构地区：[1]中国移动通信集团广东有限公司,广东广州510623

出　　处：《移动通信》2021年第3期82-85,共4页Mobile Communications

摘　　要：针对当前网络切片资源仅考虑时间域请求所导致切片资源分配不合理的问题,采用基于残差-预测框架的流量感知算法来预测网络流量,结合神经网络预测未来的网络负载状态;基于网络负载预测结果,利用VNF技术实现网络切片资源智能分配。实验表明,该方法能够提升虚拟服务功能链数据包调度的端到端时延和CPU利用率。For the unreasonable allocation issue of slicing resources caused by the current network slicing resources only considering the time domain request,this paper uses a traffic-aware algorithm based on the residual-prediction framework to predict the network traffic and the future network load status is predicted by combing a neural network.Based on the network load prediction results,the VNF technology is adopted to realize the intelligent allocation of network slicing resource.Experimental results show that the proposed method improves the end-to-end delay performance and CPU utilization of the packet scheduling in virtual service function chain.

关 键 词：流量感知 网络切片 神经网络 资源分配 

分 类 号：TN925.9[电子电信—通信与信息系统]