Enhancing DDoS Attack Detection Using Snake Optimizer With Ensemble Learning on Internet of Things Environment

A more widespread Internet of Things (IoT) device is performing a surge in cyber-attacks, with Distributed Denial of Service (DDoS) attacks posing a major risk to the reliability and availability of IoT services. DDoS attacks overwhelm target methods by flooding them with a huge volume of malicious traffic from several sources. Mitigating and identifying these attacks in IoT platforms are vital to maintaining the seamless function of IoT services and the maintenance of secret information. Feature selection (FS) is a key stage in the machine learning (ML) pipeline as it supports decreasing the data size, enhancing model outcomes, and speeding up training and inference. As part of IoT with DDoS attack detection, FS proposes to recognize a subset of IoT-related features that is optimum to represent the traffic features and distinguish between malicious and benign activities. This study designs a new DDoS attack detection using a snake optimizer with ensemble learning (DDAD-SOEL) technique on the IoT platform. The purpose of the DDAD-SOEL approach lies in the effectual and automated identification of DDoS attacks. To attain this, the DDAD-SOEL technique utilizes the SO algorithm for feature subset selection. Besides, an ensemble of three DL approaches namely long short-term memory (LSTM), bidirectional long short-term memory (BiLSTM), and deep belief network (DBN) approach. Finally, the Adadelta optimizer can be applied for the parameter tuning of the DL algorithms. The simulation value of the DDAD-SOEL methodology was tested on the benchmark database and the outcome indicates the improvements of the DDAD-SOEL methodology over other recent models in terms of distinct measures.