Immune multi agent system for intrusion prevention and self healing system implement a non-linear classification

Abstract

Artificial immune systems have recently been implemented in the field of computer security system particularly in intrusion detection and prevention systems. In this paper researchers present an approach to an intrusion prevention system (IPS) which is inspired by the Danger model of immunology. This novel approach used a multi immune agent system that implements a non-linear classification method to identify the abnormality behavior of network system. The authors look into Dendritic Cell (DC) which is a cell in Innate Immune system (IIS) as a classifier cell. Our approach takes the advantages of multi agent system, Dendritic cell, Cluster-K-Nearest-Neighbor, K-mean and Gaussion mixture methods which are give an autonomous, highly accurate and fast classifier security system. This is based on intelligent agents that exploit known functional features of the immune system and the self-healing system to detect, prevent and heal harmful or dangerous events in network systems A combination of features between the IPS and self healing (SH) mechanism to ensure continuity of the networked systems have been established.