A Metaheuristic Approach to Detecting and Mitigating DDoS Attacks in Blockchain-Integrated Deep Learning Models for IoT Applications

The Internet of Things (IoT) is a developing technology and its range of applications is satisfying among numerous consumers, as it makes everything very simple. As a concern of its huge evolution, privacy, and security are vital problems where IoT devices are always susceptible to cyberattacks. To overcome this problem, mitigation and intrusion detection are achieved which improves the safety of IoT networks. Recently, the technology of blockchain (BC) has developed as a strength for IoT-based application growth. The BC is mainly operated to resolve single-point failure (third-part dependence), security, and privacy problems of IoT applications. The combination of BC using IoT is an advantage for both society and individuals. However, a Distributed Denial of Service (DDoS) attack on extracting pool discovered the crucial mistake line between BC-permitted networks of IoT. Besides, this use makes vast data quantities. Machine Learning (ML) provides broad sovereignty in the study of big data, and abilities of decision making and so it is employed as a critical device. Therefore, this study develops a Metaheuristic Approach to Detecting and Mitigating Attacks in the Blockchain-Integrated Deep Learning (MHADMA-BCIDL) model. The designed MHADMA-BCIDL approach’s main goal is to classify and detect DDoS attacks and accomplish security in the IoT Application. To obtain this, the MHADMA-BCIDL method utilizes BC technologies to facilitate secured communication in networks of IoT, reinforcing security through decentralized and immutable data storage. For the feature selection process, the arctic tern optimization (ATO) technique is employed to decrease the data dimensionality while conserving the most relevant attributes. Besides, the MHADMA-BCIDL technique employs an attention-based convolutional neural network with bi-directional long short-term memory (CNN-BiLSTM-Attention) method for the detection and classification of attacks. For parameter tuning, the walrus optimizer (WO) model is used to fine-tune the hyperparameters of the CNN-BiLSTM-attention method. The investigational outcome study of the MHADMA-BCIDL technique is verified under BoT-IoT dataset. The experimental validation of the MHADMA-BCIDL technique portrayed a superior accuracy value of 99.32% over existing technique.