Babu Chukka Ramesh, Suneetha M, Ahmed Mohammed Altaf, Babu Palamakula Ramesh, Ishak Mohamad Khairi, Alkahtani Hend Khalid, Mostafa Samih M
Department of Electrical Communication Engineering, Vignan Institute of Information Technology, Visakhapatnam, India.
Department of Information Technology (IT), VR Siddhartha Engineering College (A), Siddhartha Academy of Higher Education (Deemed to be University), Vijayawada, India.
Sci Rep. 2024 Dec 28;14(1):30868. doi: 10.1038/s41598-024-81116-4.
Enhanced technologies of the future are gradually improving the digital landscape. Internet of Things (IoT) technology is an advanced technique that is quickly increasing owing to the development of a network of organized online devices. In today's digital era, the IoT is considered one of the most robust technologies. However, attackers can effortlessly hack the IoT devices employed to generate botnets, and it is applied to present distributed denial of service (DDoS) attacks beside networks. The DDoS attack is the foremost attack on the system that causes the complete network to go down. Thus, average consumers may need help to get the services they need from the server. The compromised or attackers IoT devices want to be perceived well in the system. So, presently, Deep Learning (DL) plays a prominent part in forecasting end-users' behaviour by extracting features and identifying the adversary in the network. This paper proposes a Synergistic Swarm Optimization and Differential Evolution with Graph Convolutional Network Cyberattack Detection and Mitigation (SSODE-GCNDM) technique in the IoT environment. The main intention of the SSODE-GCNDM method is to recognize the presence of DDoS attack behaviour in IoT platforms. Primarily, the SSODE-GCNDM technique utilizes Z-score normalization to scale the input data into a uniform format. The presented SSODE-GCNDM approach utilizes synergistic swarm optimization with a differential evolution (SSO-DE) approach for the feature selection. Moreover, the graph convolutional network (GCN) method recognizes and mitigates attacks. Finally, the presented SSODE-GCNDM technique implements the northern goshawk optimization (NGO) method to fine-tune the hyperparameters involved in the GCN method. An extensive range of experimentation analyses occur, and the outcomes are observed using numerous features. The experimental validation of the SSODE-GCNDM technique portrayed a superior accuracy value of 99.62% compared to existing approaches.
未来的增强技术正在逐步改善数字环境。物联网（IoT）技术是一种先进技术，由于有组织的在线设备网络的发展，其正在迅速发展。在当今的数字时代，物联网被认为是最强大的技术之一。然而，攻击者可以轻松地入侵用于生成僵尸网络的物联网设备，并利用其对网络进行分布式拒绝服务（DDoS）攻击。DDoS攻击是对系统最主要的攻击，会导致整个网络瘫痪。因此，普通消费者可能无法从服务器获得他们所需的服务。被入侵或受攻击者控制的物联网设备希望在系统中不被察觉。所以，目前深度学习（DL）在通过提取特征和识别网络中的对手来预测终端用户行为方面发挥着重要作用。本文提出了一种在物联网环境中基于协同群优化和差分进化与图卷积网络的网络攻击检测与缓解（SSODE-GCNDM）技术。SSODE-GCNDM方法的主要目的是识别物联网平台中DDoS攻击行为的存在。首先，SSODE-GCNDM技术利用Z分数归一化将输入数据缩放到统一格式。所提出的SSODE-GCNDM方法利用协同群优化与差分进化（SSO-DE）方法进行特征选择。此外，图卷积网络（GCN）方法用于识别和缓解攻击。最后，所提出的SSODE-GCNDM技术采用苍鹰优化（NGO）方法对GCN方法中涉及的超参数进行微调。进行了广泛的实验分析，并使用多种特征观察结果。与现有方法相比，SSODE-GCNDM技术的实验验证显示出99.62%的卓越准确率。
