Intelligent Honeypot-based IDS for Cyber Attack Detection using Generative Adversarial Networks

Abstract Cyber Intrusion is the most threatening expression in the cyber world, and it is a dangerous crime that many corporations and individuals who are a part of the Cyber World are horrified of. It results from not only a financial loss but also includes personal data which is impacted as a flooded river when the data is exposed in a data breach. Cyber Intrusion Detection System refers to a technology utilized for recognizing and notifying any illicit entry to a network system or network device. It analyses network traffic logs files maintained by a device and reports or alerts when an outsider is trying to gain access to a network. A honeypot is a technology that acts as a catchy pot of honey for an attacker. When an attacker tries to catch the pot, the Honeypot system will alert the administrator and block it. These technologies can be combined with Artificial Intelligence to automate and improve the attack prediction rate so that attackers will be prevented. A network design is proposed for the correct placement of a honeypot in the network. Datasets extracted from implementing a honeypot and KDD-99 are used in this research experimentation. The datasets are trained using Generative Adversarial Networks, a deep learning-based generative model. From the results, it is inferred that the GAN model helps in achieving enhanced performance metrics in terms of accuracy and optimum training time in both datasets.