Abstract

The IoT (Internet of Things) connects everyday things to devices that communicate with one another or with other systems across a network. DDoS (Distributed Denial of Service) attack is a significant security issue in the IoT environment. Detecting DDoS attacks is a difficult and important challenge for improving the performance of IoT technologies. Recently, a well-known recurrent neural network (RNN) called the long-short term memory (LSTM) has been used to identify DDoS attacks. However, as the LSTM model's parameters are frequently established by experience, subjectivity is significant and will have an impact on the model's capacity. In this research work, the parameters of LSTM are optimized by BCO (bacterial colony optimization) for making a more efficient DDoS detection method called BCO-LSTM. The performance of BCO-LSTM is compared with conventional LSTM and some enhanced LSTM. The investigational effects indicate that the proposed BCO-LSTM outperforms compared detectors at accurately capturing the dynamic behaviors of unknown network traffic.