Study Of Fuzzy Petri Net Attack Models Based On Reliability And Both-branch Fuzzy Set Theory

With the increasingly serious problem of network security,network attack on the defense put forward ever-increasing demands.In order to design a higher security,more comprehensive network defense system,deep and systematic understanding of attacking process is needed urgently,and the concept of Network Attack Model is presented.This paper,on the basis of previous research,establishes Network Attack Model with formalized method--Petri Net,combined both-branch fuzzy set theory and reliability theory.Because Both-branch fuzzy set applicated in attack model can expand attacking effect factors from the single espect of promoting attack to both espects of promoting and counteracting attack,it describes attacking process more actually.In this paper reliability theory is introduced to attacking model,which develops attacking model from static to dynamic.Firstly,some common limitations of previous attack model based on Petri Net is pointed out,and a new consistent Network Attack model—CBBVFP,which based on both-branch fuzzy set and fuzzy Petri Net,is established.This model extends the two basic structure"AND"and"OR",and meanwhile introduces several methods to avoid the inconsistency in this attack model.Besides,for consistency necessary,new formula of reasoning propositions truth degree and reasoning algorithm of model are presented. Secondly,from the reliability point of view,this paper makes use of fuzzy timed Petri net(FTPN)to make a dynamic reliability network attack model—DRFTPN.Comparing with previous attack models,the obvious improvement of DRFTPN is introducing the time factor,which improves static models to dynamic models.In the new model,an attacking process is described as an unrepairable complicated system,and attacking states described by Places are considered as random failure.On these bottoms the model system reliability analysis is made.Another innovation of the model is that a Petri net paths generating algorithm is constructed.Later,with the research results of DRFTPN model,another new network attack model—RBBFPRN is presented,which based on both-branch decision-making and reliability theory.The difference of RBBFPRN and DRFTPN in reliability is:an attacking state is not only random failure in RBBFPRN, but it can be repaired through a repair time after failure.This view is more according with practical fact,which is that attackers are always keep promoting attacking states.In this paper,three kinds of models and corresponding reasoning algorithms are presented and verified with examples.The verified results show the rationality of these models,and prove that these models will provide necessary references to design defense system.Above all,the paper deepens the application research of both network attack model and attacking theory,and open up new ways to study dynamic network attack models.