Research For Intrusion Detection Model And Technique Based On Artificial Immune System

The intrusion detection technique is an important method to insure the computer network security. It is becoming one of hot research topics in information security field. The operating mechanism of intrusion detection systems is naturally similar to the human immune system. The theory that the immune system can protect body from invasion provides an important approach to investigating the intrusion detection technique.In this paper, firstly, the human immune mechanisms such as special recognition, negative selection, affinity maturation, gene library and antibody diversity are analyzed. The similarity and difference between the human immune system and the artificial immune system is compared. Based on the artificial immune system, an intrusion detection model is designed to stimulate imperfect detection, negative selection, memory recognition, colonal selection, and special detector generation. The model is distributed, adaptable and robust. Secondly, the detector generation algorithm on the distributed environment is improved by using gene library evolution and hypermutation to generate detectors. Since thealgorithm considers three strategies and the equilibrium of gene library in all intrusion detection nodes, it is adapted to the distributed computing environment. Thirdly, the algorithm for realizing immune memory is designed by applying the affinity mutation algorithm and negative selection algorithm. This algorithm can get more efficient memory detectors and delete those memory detectors which may make error report in time. Finally, the detection task is distributed to all intrusion detection nodes on intrusion detection system, and a distributed parallel colonal selection algorithm is designed in use of the pipelining technique and asynchronized parallel processing mechanism. The theory analysis shows that this algorithm can get a liner speed-up.