Learning-based System for Detecting Abnormal Traffic and Host Control

Worm viruses nowadays tend not only to simply attack a host and destroy it but generate high volumes of traffic and cause network failure. This paper proposes a learning-based system for detecting abnormal traffic with a control function for individual hosts included in it for efficient protection against worm viruses and network security on a network level. The system searches, detects and learns abnormal traffic on a network level to prevent factors causing network bottleneck from affecting in advance. This paper also presents a network security management system using the ARP Spoofing attack method to efficiently control the hosts within the network.