Design of a Four-layer Model Based on Danger Theory and AIS for IDS

After a decade of research into the classical self-non-self model for immune-based intrusion detection system (IDS), a new idea has been developed by immunologists recently, which is called danger theory. Based on this new viewpoint, we present a four-layer model of artificial immune system (AIS). Also, a mechanism of reasoning with uncertainty is proposed to increase the detection accuracy. There are three merits of this novel model. First, danger theory provides a method of 'grounding' the immune response to reduce false alarms. Second, each layer of the model can work independently and interact with each other, so the efficiency of detecting is greatly improved. Last, the reasoning with uncertainty allows the model to dynamically adapt to real network traffic.