An Automatic Features Extraction Model of IDS for IOT

Abstract

The intrusion detection system (IDS) plays a significant role on the security of Internet of Things (IOT). The updated speed of attack signatures of IDS determines the detection efficiency. To improve the updated speed of intrusion detection feature database, a novel model using distributed Honeynet is proposed to capture intrusion detection features timely. The key method is to design a dynamic virtual Honeynet, which contains medium interaction honeypots and high interaction honeypots to deceive the attackers. This Honeynet can attract new attack behaviors and extract the attack signatures automatically based on the interpretable artificial intelligence. This method can effectively improve the capture speed of attack for IOT and enhance the detection capability of IDS. The experimental results show that this model is able to effectively improve the detection efficiency of IDS.