A Study For Intrusion Detection System Inspired By Danger Theory

The negative selection algorithm proposed by Forrest based on traditional immunological self-nonself theory. This algorithm works efficiently only if the self set is pure and full. However, it is difficult to get an ideal self set when applied the algorithm to real world. Due to its inherent limitation, SNS model has been challenged within immunology itself and an alternative danger theory is emerging to meet these challenges. Danger Theory states that the immune system does not discriminate on the basic of self or nonself but on the balance between the concentration of danger and safe signals with the tissue of the body.Followed by a detailed explanation of the biology information for Danger Theory and its achievement in real world, An IDS model including components and structures based on danger theory mechanism is presented and then we demonstrate the working flow process and features for each component based on their correlativity. At last, a novel algorithm is designed for sensing danger and calculating the size of danger zone.Implementation for danger theory so far has monitored danger signals modeling the mechanisms of cells which are responsible for signal defecting and processing in organism, but not taken into the concept of tissue which provides the existing environment for immune cells. This paper aims to create a contextually aware IDS by incorporating the concept of tissues. A tissue growing approach inspired from data fields is given and followed by the detailed definition of elements such as cell, signals and danger. And then a novel algorithm to sense dangers is presented. This algorithm provides a useful measure to danger signal and an efficient way to create danger zone from correlation signals. In this paper, we use data set to implement the experiment, and the experimental results show the advantages of danger theory with high adaptability and efficiency.