Multi-Layered Moder Inspired By Immune Principle To Network Fault Detection

The rapid development of computer networks brings a great convenience for people, however, it also poses challenges to people, with the increasing complex network structure, more and more sophisticated equipment, the possibility of network failure occurs more and more, and even result in paralysis of part of the network. Therefore, Network managers are facing the problem is how to effectively monitor the network insecurity factors or network faults. When the network or component can not achieve the required performance indicators, network managers must quickly find faults.In order to reduce the fault detection rate and improve the self-adaptive capability in network fault detection, an artificial immune mechanism which is inspired by multi-layer defense of a biological immune system is proposed to perform network fault detection. The immune model is composed of three parts: the inherent detection layer, fuzzy judgment layer and adaptive detection layer. Dendritic cells can influence the reaction of coordinating T-cells, which can be activate or tolerate so that induce adaptive immune responses and affirm the type of adaptive response. Inherent detection layer and fuzzy judgment layer interact with each other to reduce the error detection rate, while the adaptive detection layer is capable of learning unknown fault patterns.In this paper, it introduces the innate and adaptive immune system of biological immune system. The innate immune system plays an important defense mechanism in the biological immune system. Artificial immune system should not ignore the important role of the innate immune cells. The Multi-layered immune model draws on the natural immune system, With a powerful information processing capabilities, including self/non-self pattern recognition, the adaptive immune response is affected by the innate immune cells, the clonal selection algorithm and the distributed features of biological immune system.In the fuzzy layer, dendritic cells is an antigen presenting cells, which is considered a natural detector can accurately gather a variety of signals around and react to them. Established at this level collaboration diagram between devices, when a device occurs abnormal phenomenon, Collaborative signal is calculated based on collaboration diagram between devices.In the adaptive layer, through improved clonal selection algorithm to generate unknown fault detectors. At the same time, mature detectors are set life cycle, to ensure the effectiveness of detectors so that the system has the best recognition ability.Finally, as an example of the campus network, select the part of the common interface failures, to verify validity of the model.