CNN-LSTM: A Deep Learning Model to Detect Botnet Attacks in Internet of Things

he Internet of Things has seen massive growth in recent years. This growth can be primarily linked to the advancement in technology with special attribution to the 5G technology. IoTs are present virtually everywhere. We can use them not just in our homes but in critical infrastructure and industries. An increase in this popularity has also led to IoT being targeted by cybercriminals. Since IoTs are not designed with security as their primary objective, they are ravaged by various vulnerabilities and threats. Botnets amount to a major chunk of these threats in IoT. Mirai and BASHLITE are some of the prominent examples of such threats. With the increase in popularity of IoT, there also arises a need to make IoT more secure. Manually mitigating threats from IoT can prove to be a herculean task. Considering the effectiveness of artificial intelligence techniques in automating manual tasks with degree of accuracy and low error rate. We propose a deep learning-based hybrid model. This model can be employed to detect botnet threats by analyzing the network traffic emanating from an IoT device. This proposed model is built upon the CTU-13 dataset and outperforms traditional deep learning models.