基于ResNet18的SRv6网络流量异常检测研究

doi:10.11959/j.issn.1000-0801.DXKX260009

Chinese

您当前的位置：

首页 >

文章列表页 >

基于ResNet18的SRv6网络流量异常检测研究

更新时间：2026-05-07

- 基于ResNet18的SRv6网络流量异常检测研究
- Research on SRv6 Network Traffic Anomaly Detection Based on ResNet18
- 电信科学 2026年
- 作者机构：
- 作者简介：
- 基金信息：
  
  国家自然科学基金资助项目;浙江省教育厅科学基金资助项目
- DOI：10.11959/j.issn.1000-0801.DXKX260009
  中图分类号： TP393.08
- 收稿：2026-01-06，
  
  修回：2026-03-02，
  
  录用：2026-04-09，
- 稿件说明：
移动端阅览
基于ResNet18的SRv6网络流量异常检测研究[J/OL]. 电信科学, 2026.

Research on SRv6 Network Traffic Anomaly Detection Based on ResNet18[J/OL]. Telecommunications Science, 2026.
基于ResNet18的SRv6网络流量异常检测研究[J/OL]. 电信科学, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260009.

Research on SRv6 Network Traffic Anomaly Detection Based on ResNet18[J/OL]. Telecommunications Science, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260009.

摘要

随着基于 IPv6 的段路由（Segment Routing over IPv6，SRv6）在新一代网络中的广泛部署，其可编程性和灵活性显著提升了网络性能，但也带来了新的安全挑战。传统的基于特征规则或阈值的方法在应对复杂、多样化的攻击流量时往往表现不佳，尤其难以识别针对 SRH（Segment Routing Header）的异常行为。因此，如何在 SRv6 环境下实现高效、准确的异常流量检测，成为亟待解决的问题。为此，本文提出了一种基于深度学习的 SRv6 网络流量异常检测方法。实验结果表明，本文方法在准确率、召回率和 F1-score 等指标上均优于传统机器学习算法和部分主流深度学习模型，能够有效识别多类攻击流量。本文的主要创新点包括：（1）构建了“异常类型—可观测特征—检测指标”三层映射机制，系统性刻画SRv6协议特有异常行为的可感知特征；（2）提出一种面向SRv6场景的流量图像化表示方法，并结合残差网络实现高维特征的自适应提取；（3）在缺乏公开SRv6异常数据集的条件下，设计了基于协议规范与威胁模型的特征扩展机制，实现了SRv6场景下的验证性实验框架。在SRv6场景下，该方法为面向可编程网络的安全防护提供了一种可扩展的技术路径。

Abstract

With the widespread deployment of Segment Routing over IPv6 (SRv6) in next-generation networks

its programmability and flexibility have significantly improved network performance

but have also brought new security challenges. Traditional methods based on feature rules or thresholds often perform poorly when dealing with complex and diverse attack traffic

especially struggling to identify anomalous behavior targeting the SRH (Segment Routing Header). Therefore

how to achieve efficient and accurate anomaly traffic detection in the SRv6 environment has become an urgent problem to be solved. To this end

this paper proposes a deep learning-based method for anomaly detection of SRv6 network traffic. Experimental results show that the proposed method outperforms traditional machine learning algorithms and some mainstream deep learning models in terms of accuracy

recall

and F1-score

and can effectively identify multiple types of attack traffic. The main innovations of this paper can be summarized as follows: (1) A three-layer mapping mechanism of "anomaly type - observable features - detection index" is constructed to systematically characterize the perceptible features of the unique abnormal behavior of the SRv6 protocol; (2) A traffic image representation method for SRv6 scenarios is proposed

and adaptive extraction of high-dimensional features is achieved by combining residual networks; (3) In the absence of publicly available SRv6 anomaly datasets

a feature extension mechanism based on protocol specifications and threat models is designed to realize a verification experimental framework for SRv6 scenarios. In the SRv6 scenario

this method provides a scalable technical path for security protection of programmable networks.

关键词

Keywords

references

Stefano Previdi , Clarence Filsfils , et al . 2018 . IPv6 Segment Routing Header (SRH) . Internet-Draft draft-ietf-6man-segment-routing-header-14 .

Mo Z , Long B . An Overview of SRv6 Standardization and Application towards 5G-Advanced and 6G [C ] // 2022 IEEE 5th International Conference on Computer and Communication Engineering Technology (CCET) . IEEE , 2022 : 266 - 270 .

Hwang R H , Peng M C , Huang C W , et al . An unsupervised deep learning model for early network traffic anomaly detection [J ] . IEEE Access , 2020 , 8 : 30387 - 30399 .

Wang J , Yu P , Xu S , et al . High-Accuracy Fault Diagnosis for SRv6 TE Policy in Computer Power Networks [C ] // 2024 IEEE/CIC International Conference on Communications in China (ICCC Workshops) . IEEE , 2024 : 60 - 65 .

Liu B . Supervised learning [M ] // Web Data Mining: Exploring Hyperlinks, Contents, and Usage Data . Berlin, Heidelberg : Springer Berlin Heidelberg , 2011 : 63 - 132 .

James G , Witten D , Hastie T , et al . Unsupervised learning [M ] // An introduction to statistical learning: with applications in Python . Cham : Springer International Publishing , 2023 : 503 - 556 .

Van Engelen J E , Hoos H H . A survey on semi-supervised learning [J ] . Machine learning , 2020 , 109 ( 2 ): 373 - 440 .

M. Al-Qatf , Y. Lasheng , M. Al-Habib and K . Al-Sabahi , " Deep Learning Approach Combining Sparse Autoencoder With SVM for Network Intrusion Detection ," in IEEE Access , vol. 6 , pp. 52843 - 52856 , 2018, doi: 10.1109/ACCESS.2018.2869577 http://dx.doi.org/10.1109/ACCESS.2018.2869577 .

W . Wan g et al ., " HAST-IDS: Learning Hierarchical Spatial-Temporal Features Using Deep Neural Networks to Improve Intrusion Detection ," in IEEE Access , vol. 6 , pp. 1792 - 1806 , 2018, doi: 10.1109/ACCESS.2017.2780250 http://dx.doi.org/10.1109/ACCESS.2017.2780250 .

He K , Zhang X , Ren S , et al . Delving deep into rectifiers: Surpassing human-level performance on imagenet classification [C ] // Proceedings of the IEEE international conference on computer vision . 2015 : 1026 - 1034 .

Srivastava N , Hinton G , Krizhevsky A , et al . Dropout: a simple way to prevent neural networks from overfitting [J ] . The journal of machine learning research , 2014 , 15 ( 1 ): 1929 - 1958 .

王强 , 杨宏 . 基于SRv6技术实现高可靠网络传输研究 [J ] . 通信技术 , 2025 , 58 ( 01 ): 39 - 49 .

Wang S , Balarezo J F , Kandeepan S , et al . Machine learning in network anomaly detection: A survey [J ] . IEEe Access , 2021 , 9 : 152379 - 152396 .

Szegedy C , Ioffe S , Vanhoucke V , et al . Inception-v4, inception-resnet and the impact of residual connections on learning [C ] // Proceedings of the AAAI conference on artificial intelligence . 2017 , 31(1) .

He K , Zhang X , Ren S , et al . Deep residual learning for image recognition [C ] // Proceedings of the IEEE conference on computer vision and pattern recognition . 2016 : 770 - 778 .

Floristean G R , Udrea A . Detection and Classification of Anomalies in IP Communications Networks [J ] . Journal of Control Engineering and Applied Informatics , 2021 , 23 ( 4 ): 25 - 32 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于GPT网络的日志文本故障关联规则挖掘研究

基于深度学习的软件定义网络应用策略冲突检测方法

基于CVAE-ReS-GRU算法的电网调控系统异常检测方法研究

基于SDL-YOLO的红外图像人车检测算法