一种基于图神经网络的SDN路由性能预测模型  

A Prediction Model of SDN Routing Performance Based on Graph Neural Network

作　　者：车向北 康文倩 邓彬 杨柯涵 李剑[2] CHE Xiang-bei;KANG Wen-qian;DENG Bing;YANG Ke-han;LI Jian(Shenzhen Power Supply Bureau Co.,Ltd.,Shenzhen,Guangdong 518000,China;School of Computer Science,Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]深圳供电局有限公司,广东深圳518000 [2]北京邮电大学计算机学院,北京100876 [3]深圳供电局有限公司系统运行部,广东深圳518000

出　　处：《电子学报》2021年第3期484-491,共8页Acta Electronica Sinica

基　　金：国家自然科学基金(No.U1636106);北京市自然科学基金(No.4182006)。

摘　　要：软件定义网络作为未来网络架构的发展方向,通过分离数据平面与控制平面高效设定路由方案.而在路由方案的优化过程中,准确预估给定路由方案下的网络性能是其关键.本文基于图神经网络建模网络中物理链路与路由方案路径的关系,在给定的路由方案与网络流量下对网络中的各项端到端性能指标(如延迟、抖动)进行准确预估,以辅助优化路由方案.本文基于OMNeT++来生成数据并进行实验,实验结果表明本文提出的模型能够针对延迟抖动等端到端性能指标进行准确预估,预估平均相对误差不超过4.1%.实验也对比了传统最短路径路由算法与基于该预测模型给出的最优路由方案下的端到端性能,相比传统最短路径路由算法,平均延迟和平均抖动分别降低了19.8%和33.52%,最大延迟和最大抖动降低了36.18%和35.45%.As the development direction of future network architectures,Software Defined Networks can efficiently set routing schemes by separating the data plane and the control plane.In the process of optimizing a routing scheme,it is the key to accurately predict the network performance under a given routing scheme.This paper uses graph neural networks to model the relationship between physical links and routing scheme paths,so that the model can predict various end-to-end performance indicators(such as delay and jitter)in the network under a given routing scheme and network traffic.This paper uses OMNeT++to generate datasets.The experimental results show that the model proposed in this paper can accurately predict end-to-end performance indicators such as delay and jitter.The average relative error of the estimate does not exceed 4.1%.The experiment also compares the end-to-end performance of the traditional shortest path routing algorithm with the optimal routing scheme based on the prediction model proposed in this paper.Compared to the traditional shortest path routing algorithm,the average delay and average jitter are reduced by 19.8%and 33.52%,and the maximum delay and maximum jitter are reduced by 36.18%and 35.45%.

关 键 词：软件定义网络 端到端性能预测 图神经网络 SDN路由优化 

分 类 号：TP393.0[自动化与计算机技术—计算机应用技术]