Algorithms Based On Danger Theory And Its Application On Network Security

With the further research and achievements made in Artificial ImmuneSystem (AIS) these years, a newly-introduced immune system, the DangerTheory, has challenged against the basic theories of modern immune based onthe traditional SNS. It has broken the concept of self tolerance and smashed thetrammels of those traditional theories by changing the way that we looked at theconcept of space. Danger signals introduced by Danger Theory are able todemonstrate and process a wealth of data effectively. Therefore, this newlyintroduced immune theory has started a brand-new vision for the research ofAIS.Inspired by Danger Theory and based on the previous theories, algorithmsbased on Danger Theory and its application into network security has beenstudied in this paper. And the results are the following.1,AIS based on Danger Theory and the traditional theories of AIS are comparedin terms of immunology theory and algorithms; the limit of existing SNS isanalyzed; and the advantages of AIS based on Danger Theory are expatiated,then DIA is proposed.2,The traditional immune algorithm is not perfect in its application into networksecurity, such as its high computation complexity and low self-adaptability.Therefore, Danger Theory is introduced, from which principles and structuresconcerned have been concluded, then the feasibility of their application intoanomaly detection is analyzed and a model of anomaly detection for IntrusionDetection System (IDS) and a forecasting system for unknown Internet worms based on Danger Theory have been proposed. At last, combined with clustermethod, algorithm concemed is put forward.3,The experiment on DIA and the forecasting system for unknown Internetworms is made and the result is analyzed.4,Finally, the paper describes the respective of the application of algorithmsbased on Danger Theory into network security, and indicates that the work needto be further studied.