Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (2): 505-517.DOI: 10.11772/j.issn.1001-9081.2025020200

• Network and communications • Previous Articles    

Performance evaluation method for deterministic networks based on complex-enhanced attention graph neural network

Junrui WU1, Jiangchuan YANG2, Haisheng YU3, Sai ZOU4, Wenyong WANG1()   

  1. 1.School of Computer Science and Engineering (School of Cyber Security),University of Electronic Science and Technology of China,Chengdu Sichuan 611731,China
    2.Product Research and Development Department,Motorcomm (Shanghai) Electronic Technology Company Limited,Shanghai 201206,China
    3.China Internet Network Information Center,Beijing 100070,China
    4.College of Big Data and Information Engineering,Guizhou University,Guiyang Guizhou 550025,China
  • Received:2025-03-03 Revised:2025-06-10 Accepted:2025-06-11 Online:2025-06-12 Published:2026-02-10
  • Contact: Wenyong WANG
  • About author:WU Junrui, born in 1993, Ph. D. candidate. His research interests include computer network architecture, cyber security.
    YANG Jiangchuan, born in 1998, M. S. His research interests include deterministic network transmission, computer network architecture.
    YU Haisheng, born in 1983, Ph. D. His research interests include deterministic network transmission, next-generation Internet.
    ZOU Sai, born in 1981, Ph. D., professor. His research interests include next-generation Internet, biomedical informatics, smart education.
    WANG Wenyong, born in 1967, Ph. D., professor. His research interests include computer networks, cyber security. Email:wangwy@uestc.edu.cn
  • Supported by:
    National Natural Science Foundation of China-Civil Aviation Joint Research Fund Key Project(U033212);NSFC Original Exploration Program(62250067);Natural Science Foundation of Sichuan Province(2024NSFSC0473);Fundamental Research Funds for the Central Universities(ZYGX2024Z009)

基于复增强注意力机制图神经网络的确定性网络性能评估方法

吴俊锐1, 杨江川2, 喻海生3, 邹赛4, 汪文勇1()   

  1. 1.电子科技大学 计算机科学与工程学院(网络空间安全学院),成都 611731
    2.裕太微(上海)电子有限公司 产品研发部,上海 201206
    3.中国互联网络信息中心,北京 100070
    4.贵州大学 大数据与信息工程学院,贵阳 550025
  • 通讯作者: 汪文勇
  • 作者简介:吴俊锐(1993—),男,四川安岳人,博士研究生,主要研究方向:计算机网络体系结构、网络安全
    杨江川(1998—),男,上海人,硕士,主要研究方向:确定性网络传输、计算机网络体系结构
    喻海生(1983—),男,安徽庐江人,博士,主要研究方向:确定性网络传输、下一代互联网
    邹赛(1981—),男,湖南衡阳人,教授,博士,主要研究方向:下一代互联网、生物信息化、智慧教育
    汪文勇(1967—),男,四川简阳人,教授,博士,CCF杰出会员,主要研究方向:计算机网络、网络安全。 Email:wangwy@uestc.edu.cn
  • 基金资助:
    国家自然科学基金-民航联合基金重点项目(U033212);国家自然科学基金原创探索项目(62250067);四川省自然科学基金资助项目(2024NSFSC0473);中央高校项目(ZYGX2024Z009)

Abstract:

Deterministic network is a key focus and hotspot in the current development of the Internet. How to evaluate the performance of non-deterministic traffic in deterministic network systems remains a challenge both in theory and engineering. Although traditional methods, such as modeling network traffic arrival distribution using queuing theory, employing network calculus analysis to establish upper and lower bounds of network behavior, and applying machine learning and deep learning algorithms to predict network performance trends from large-scale historical data statistically, can address this issue partially, they still suffer from poor accuracy and performance facing complex deterministic network systems. Therefore, a Complex-Enhanced Attention Graph Neural Network (CEA-GNN)-based method was proposed for deterministic network performance evaluation. In the method, the gating property of deterministic network systems was utilized fully, and an attention mechanism was employed to deliver critical information to the Graph Neural Network (GNN), thereby improving the evaluation accuracy. At the same time, the relevant information was extracted from both the graph spatial domain and the complex frequency domain to update the GNN, thereby enhancing performance of the evaluation model. Experimental results from experiments conducted on the National Science Foundation Network (NSFNet) topology indicate that compared to the RouteNet-Fermi evaluation method, the proposed method reduces the Mean Absolute Error (MAE) of non-deterministic traffic latency prediction in deterministic networks by 87.4%, decreases the MAE of packet loss rate prediction by 12.7%, and reduces the average processing time per flow by 64.4%.

Key words: deterministic network, performance evaluation, Graph Neural Network (GNN), complex-enhanced attention mechanism, hidden information extraction

摘要:

确定性网络是当前互联网发展的重点和热点,如何评估确定性网络系统中非确定性流量的性能在理论和工程上仍是一个挑战。尽管传统方法,如使用排队论建模网络流量到达分布、采取网络演算分析建立网络行为的上下界以及应用机器学习和深度学习算法从大量历史数据中统计推测网络性能趋势可以部分解决这一问题,但面对复杂的确定性网络系统,它们依然存在精度和性能较差的缺陷。因此,提出一种基于复增强注意力机制图神经网络(CEA-GNN)的确定性网络性能评估方法。该方法充分利用确定性网络系统的门控属性,并使用注意力机制将关键信息传递给图神经网络(GNN),从而提高评估精度;同时,从图形空间域和复频域中提取相关信息更新GNN,以提升评估模型的性能。在美国国家科学基金会网络提供的网络拓扑上的实验结果表明,相较于RouteNet-Fermi评估方法,所提方法对确定性网络中非确定性流量时延预测的平均绝对误差(MAE)减小了87.4%,丢包率预测的MAE减小了12.7%,每条流平均处理时间缩减了64.4%。

关键词: 确定性网络, 性能评估, 图神经网络, 复增强注意力机制, 隐信息提取

CLC Number: