基于时空图神经网络的应用层DDoS攻击检测方法

doi:10.3969/j.issn.1671-1122.2024.04.002

摘要/Abstract

摘要：

分布式拒绝服务（Distributed Denial of Service，DDoS）攻击已经成为网络安全的主要威胁之一，其中应用层DDoS攻击是主要的攻击手段。应用层DDoS攻击是针对具体应用服务的攻击，其在网络层行为表现正常，传统安全设备无法有效抵御。同时，现有的针对应用层DDoS攻击的检测方法检测能力不足，难以适应攻击模式的变化。为此，文章提出一种基于时空图神经网络（Spatio-Temporal Graph Neural Network，STGNN）的应用层DDoS攻击检测方法，利用应用层服务的特征，从应用层数据和应用层协议交互信息出发，引入注意力机制并结合多个GraphSAGE层，学习不同时间窗口下的实体交互模式，进而计算检测流量与正常流量的偏差，完成攻击检测。该方法仅利用时间、源IP、目的IP、通信频率、平均数据包大小5维数据便可有效识别应用层DDoS攻击。由实验结果可知，该方法在攻击样本数量较少的情况下，与对比方法相比可获得较高的Recall和F1分数。

关键词: DDoS攻击, 时空图神经网络, 异常检测, 注意力机制

Abstract:

Distributed denial of service (DDoS) attacks have emerged as one of the principal threats to cybersecurity, among which application layer DDoS attacks stand as a primary mode of assault. Application layer DDoS attacks target specific application services and exhibit normal behavior at the network layer, rendering traditional security devices ineffective against them. Moreover, existing detection methods for application layer DDoS attacks are insufficient in detection capability and struggle to adapt to the changing patterns of attacks. In response, this paper proposed a detection method for application layer DDoS attacks based on spatio-temporal graph neural network (STGNN). This method utilized the characteristics of application layer services, starting from application layer data and protocol interaction information. It introduced an attention mechanism and combined multiple GraphSAGE layers to learn the patterns of entity interactions across different time windows. Consequently, it calculated the deviation between the detection traffic and normal traffic to accomplish attack detection. The method effectively identified application layer DDoS attacks using only five dimensional data, including time, source IP, destination IP, communication frequency, and average packet size. According to the experimental results, this method achieves higher Recall and F1 scores compared to benchmark methods, even with a smaller number of attack samples.

Key words: DDoS attacks, spatio-temporal graph neural network, anomaly detection, attention mechanism

中图分类号:

TP309

王健, 陈琳, 王凯崙, 刘吉强. 基于时空图神经网络的应用层DDoS攻击检测方法[J]. 信息网络安全, 2024, 24(4): 509-519.

WANG Jian, CHEN Lin, WANG Kailun, LIU Jiqiang. Application Layer DDoS Detection Method Based on Spatio-Temporal Graph Neural Network[J]. Netinfo Security, 2024, 24(4): 509-519.

图/表 10

图1

图2

表1

图3

图4

表2

图5

图6

图7

图8

参考文献 32

[1]	JANG Yingzhao, CHEN Lei, YAN Qiao. Distributed Denial of Service Attack Detection Algorithm Based on Two-Channel Feature Fusion[J]. Netinfo Security, 2023, 23(7): 86-97.
	蒋英肇, 陈雷, 闫巧. 基于双通道特征融合的分布式拒绝服务攻击检测算法[J]. 信息网络安全, 2023, 23(7):86-97.
[2]	China Unicom Digital Science & Technology. Global DDoS Attack Threat Report 2023[EB/OL]. (2024-01-04)[2024-01-29]. https://www.199it.com/archives/1669730.html.
	联通数科. 2023年全球DDoS攻击威胁报告[EB/OL]. (2024-01-04) [2024-01-29]. https://www.199it.com/archives/1669730.html.
[3]	FAN Mingyu, LI Ke. A Method to Distinguish DDoS Attack Types Based on RNN[J]. Netinfo Security, 2022, 22(7): 1-8.
	范明钰, 李珂. 一种基于RNN区分DDoS 攻击类型的方法[J]. 信息网络安全, 2022, 22(7):1-8.
[4]	StormWall. Q1 2023 in Review: DDoS Attacks Report by Storm-Wall[EB/OL]. (2023-04-17)[2024-01-29]. https://stormwall.network/ddos-report-stormwall-q-1-2023.
[5]	MITTAL M, KUMAR K, BEHAL S. Deep Learning Approaches for Detecting DDoS Attacks: A Systematic Review[J]. Soft Computing, 2023, 27(18): 13039-13075.
[6]	DORIGUZZI-CORIN R, SIRACUSA D. FLAD: Adaptive Federated Learning for DDoS Attack Detection[EB/OL]. (2022-05-13)[2024-01-20]. https://arxiv.org/abs/2205.06661.
[7]	WANG Zhi, ZHANG Hao, GU Jianjun. A Hybrid Method of Joint Entropy and Multiple Clustering Based DDoS Detection in SDN[J]. Netinfo Security, 2023, 23(10): 1-7.
	王智, 张浩, 顾建军. SDN网络中基于联合熵与多重聚类的DDoS攻击检测[J]. 信息网络安全, 2023, 23(10):1-7.
[8]	KEMP C, CALVERT C, KHOSHGOFTAAR T M, et al. An Approach to Application-Layer DoS Detection[EB/OL]. (2023-02-13)[2024-01-20]. https://journalofbigdata.springeropen.com/articles/10.1186/s40537-023-00699-3.
[9]	JIANG Jianguo, YU Qian, YU Min, et al. ALDD: A Hybrid Traffic-User Behavior Detection Method for Application Layer DDoS[C]// IEEE. 2018 17th IEEE International Conference on Trust, Security and Privacy in Computing and Communications/12th IEEE International Conference on Big Data Science and Engineering. New York: IEEE, 2018: 1565-1569.
[10]	SINGH K J, DE T. MLP-GA Based Algorithm to Detect Application Layer DDoS Attack[J]. Journal of Information Security and Applications, 2017, 36: 145-153.
[11]	ZHOU Lu, SOOD K, XIANG Yong. ERM: An Accurate Approach to Detect DDoS Attacks Using Entropy Rate Measurement[J]. IEEE Communications Letters, 2019, 23(10): 1700-1703.
[12]	XIE Yi, YU Shunzheng. A Large-Scale Hidden Semi-Markov Model for Anomaly Detection on User Browsing Behaviors[J]. IEEE/ACM Transactions on Networking, 2008, 17(1): 54-65.
[13]	PRASEED A, THILAGAM P S. Modelling Behavioural Dynamics for Asymmetric Application Layer DDoS Detection[J]. IEEE Transactions on Information Forensics and Security, 2020, 16: 617-626.
[14]	SARAVANAN A, BAMA S S, KADRY S, et al. A New Framework to Alleviate DDoS Vulnerabilities in Cloud Computing[J]. International Journal of Electrical & Computer Engineering, 2019, 9(5): 4163-4175.
[15]	GULIHAR P, GUPTA B B. Cooperative Mechanisms for Defending Distributed Denial of Service(DDoS) Attacks[EB/OL]. (2020-01-01)[2024-01-29]. https://link.springer.com/chapter/10.1007/978-3-030-22277-2_16.
[16]	SAKHARE S R, PATIL V D. Implementation of Captcha Mechanisms Using Deep Learning to Prevent Automated Bot Attacks[J]. Research Journal of Computer Systems and Engineering, 2023, 4(2): 1-15.
[17]	BHUYAN M H, BHATTACHARYYA D K, KALITA J K. An Empirical Evaluation of Information Metrics for Low-Rate and High-Rate DDoS Attack Detection[J]. Pattern Recognition Letters, 2015, 51: 1-7.
[18]	FORTUNATI S, GINI F, GRECO M S, et al. An Improvement of the State-of-the-Art Covariance-Based Methods for Statistical Anomaly Detection Algorithms[J]. Signal, Image and Video Processing, 2016, 10: 687-694.
[19]	DORIGUZZI-CORIN R, MILLAR S, SCOTT-HAYWARD S, et al. LUCID: A Practical, Lightweight Deep Learning Solution for DDoS Attack Detection[J]. IEEE Transactions on Network and Service Management, 2020, 17(2): 876-889.
[20]	LIU Xinqian, REN Jiadong, HE Haitao, et al. A Fast All-Packets-Based DDoS Attack Detection Approach Based on Network Graph and Graph Kernel[EB/OL]. (2021-07-01)[2024-01-20]. https://www.sciencedirect.com/science/article/pii/S1084804521001016.
[21]	KIPF T N, WELLING M. Semi-Supervised Classification with Graph Convolutional Networks[EB/OL]. (2016-09-09)[2024-01-25]. https://www.semanticscholar.org/paper/Semi-Supervised-Classification-with-Graph-Networks-Kipf-Welling/36eff562f65125511b5dfab68ce7f7a943c27478.
[22]	HAMILTON W, YING Zhitao, LESKOVEC J. Inductive Representation Learning on Large Graphs[J]. Advances in Neural Information Processing Systems, 2017, 30: 1025-1035.
[23]	VELIČKOVIĆ P, CUCURULL G, CASANOVA A, et al. Graph Attention Networks[EB/OL]. (2017-10-30)[2024-01-25]. https://arxiv.org/abs/1710.10903.
[24]	YU Bing, YIN Haoteng, ZHU Zhanxing. Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecasting[C]// ACM. The 27th International Joint Conference on Artificial Intelligence. New York: ACM, 2018: 3634-3640.
[25]	HE Xiangnan, DENG Kuan, WANG Xiang, et al. LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation[C]// ACM. 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2020: 639-648.
[26]	SHI Chence, XU Minkai, GUO Hongyu, et al. A Graph to Graphs Framework for Retrosynthesis Prediction[C]// ACM. 37th International Conference on Machine Learning. New York: ACM, 2020: 8818-8827.
[27]	LIN Honggang, ZHANG Yunli, GUO Nanxin, et al. P2P Botnet Detection Method Based on Graph Neural Network[J]. Advanced Engineering Sciences, 2022, 54(2): 65-72.
	林宏刚, 张运理, 郭楠馨, 等. 基于图神经网络的P2P僵尸网络检测方法[J]. 工程科学与技术, 2022, 54(2):65-72.
[28]	XIE Qian, HUANG Zheng, GUO Jie, et al. Spatio-Temporal Graph Convolutional Networks for DDoS Attack Detecting[C]//ACM. Machine Learning for Cyber Security:Third International Conference. New York: ACM, 2020: 153-159.
[29]	JAZI H H, GONZALEZ H, STAKHANOVA N, et al. Detecting HTTP-Based Application Layer DoS Attacks on Web Servers in the Presence of Sampling[J]. Computer Networks, 2017, 121: 25-36.
[30]	PÉREZ-DÍAZ J A, VALDOVINOS I A, CHOO K K R, et al. A Flexible SDN-Based Architecture for Identifying and Mitigating Low-Rate DDoS Attacks Using Machine Learning[J]. IEEE Access, 2020, 8: 155859-155872.
[31]	LIMA FILHO F S, SILVEIRA F A F, de MEDEIROS BRITO JUNIOR A, et al. Smart Detection: An Online Approach for DoS/DDoS Attack Detection Using Machine Learning[EB/OL]. (2019-10-13)[2024-02-11]. https://doi.org/10.1155/2019/1574749.
[32]	BAKAR R A, HUANG Xin, JAVED M S, et al. An Intelligent Agent-Based Detection System for DDoS Attacks Using Automatic Feature Extraction and Selection[EB/OL]. (2023-03-22)[2024-01-30]. https://www.mdpi.com/1424-8220/23/6/3333.

DDoS攻击类型	被攻击IP	攻击起始时间偏移	攻击持续时间
slowbody2	75.127.97.72	53分钟	11:54:06~12:08:00
slowread	75.127.97.72	1小时58 分钟	12:59:06~13:05:50
ddossim	75.127.97.72	2小时22分钟	13:23:06~13:25:13
goldeneye	75.127.97.72	2小时50分钟	13:51:23~13:55:13
slowheaders	74.63.40.21	2小时57分钟	13:58:06~14:07:42
rudy	75.127.97.72	3小时8分钟	14:09:06~14:10:11
ddossim	97.74.144.108	3小时28分钟	14:29:06~14:32:15
rudy	208.113.162.153	3小时29分钟	14:30:06~14:31:11
hulk	69.84.133.138	4小时38分钟	15:39:06~15:44:25
slowheaders	67.220.214.50	6小时	17:01:06~17:10:42
goldeneye	97.74.144.108	7小时6分钟	18:07:23~18:11:13
slowbody2	69.192.24.88	8小时13分钟	19:14:06~19:28:00
slowbody2	97.74.144.108	9小时3分钟	20:04:06~20:18:00
slowbody2	203.73.24.75	9小时9分钟	20:10:06~20:24:02
rudy	97.74.144.108	9小时20分钟	20:21:06~20:22:11
slowread	74.55.1.4	11小时2分钟	22:03:06~22:09:50
slowheaders	97.74.104.201	11小时27分钟	22:28:06~22:37:42
hulk	74.55.1.4	13小时33分钟	24:34:06~24:44:14
hulk	69.192.24.88	13小时47分钟	24:48:06~24:56:37
slowloris	97.74.144.108	15小时20分钟	次日2:21:06~2:23:36
slowheaders	97.74.144.108	15小时47分钟	次日2:48:06~2:57:42
slowloris	75.127.97.72	16小时33分钟	次日3:34:06~3:36:36
slowheaders	75.127.97.72	17小时13分钟	次日4:14:06~4:23:42
goldeneye	69.192.24.88	19小时23分钟	次日6:24:23~6:28:13
hulk	75.127.97.72	19小时25分钟	次日6:26:06~6:31:25
rudy	74.55.1.4	20小时59分钟	次日8:00:06~8:01:11

方法	Accuracy	Precision	Recall	FAR	F1分数
本文方法	0.975	0.965	0.995	0.063	0.986
PéREZ-DíAZ^[30]等人	0.951	0.955	0.945	0.039	0.950
LIMA FILHO^[31]等人	N/A	0.999	0.936	0.004	N/A
BAKAR^[32]等人	0.936	0.999	N/A	0.040	N/A