Networked Software Anomaly Source Location And Analysis The Credibility

Since Internet entering the commercial application,along with the rapid development of computer and communication technology,the Internet has penetrated into all aspects of our work and daily life.Networked software is a class of intensive mixing system,with the Internet as the medium,topological structure and behavior can dynamic evolution.The generation and diffusion of abnormal behavior will bring great threat to the security of the system.Therefore,it is necessary to conduct in-depth study of the abnormal behavior.The main research contents of this paper are the following two aspects:(1)Reverse finding the source of abnormal behavior based on maximum likelihood estimation algorithm is proposed in this paper to locate the anomalous propagation source.In the light of rapid propagation of abnormal behavior could bring huge security risks to the system,using the RLML algorithm to locate the abnormal source.Within a certain period of time,local structures of the system has changed,the overall structure of the networked system has no big change.We combing with the regular of abnormal behavior propagation,starting from the current infection node,reverse search and restore path,calculate the maximum likelihood probability and matching the best spread path of the abnormal behavior,the focal point of best paths is the source of the abnormal behavior.The validity and effectiveness of the proposed algorithm are verified by simulation experiment and case analysis.(2)This paper makes a deep research of abnormal behavior from formation level and system level,a method for evaluating impact strength of abnormal node is proposed.First of all,the harm caused by abnormal behavior is analyzed,then the active frequency of components in the system is considered,and the two factors are combined to explore the impact of abnormal behavior to the system,it is considered that the bigger the influence bring to the networked software system of the abnormal node,the worse the credibility it is.Simulation experiment and case analysis show that research credibility of abnormal behavior can provide some help for people to assess the stability of the system andreduce the loss caused by abnormal behavior.