Study Of An Attack Model Based On Attack Graph And Petri Net

With the development and popularization of Internet, the world has entered the Internet age. Internet has characters of opening and sharing, it has expanded to an unprecedented degree of the scale, but followed by network security issues. Network security attack and defense is closely related to each other. Further investigate, attack theory and technology can protect network security of information system effectively.Firstly, according to the possibility that an attacker can be detected during a network attack, a synthetic analysis is made about an attack process both in success rate and failure rate by introducing both-branch fuzzy sets to an Attack Graph model; the description of network attack is refined into network component level, not only components but also connections among components are fuzzed, which makes the model to be closer to real situation. Finally a new network attack model named BBFCAG (both-branch fuzzy components-based Attack Graph model) is put forward, and a strategy that restrains the number of attack steps is adapted to reduce the scale of the model. Through an experiment based on modeling a real network attack, the results show that the model can make a comprehensive analysis to network attack, and produce Attack Graph, attack path and attack success rate (or failure rate).Secondly, this paper transforms Attack Graph into Petri net and defines an extended perti net which is named EPN by adding another two types of token. Defines a new concept of attack-cost to instead the concept of attack complexity in Petri net. Use FAT(first accomplished token) to produce the best attack path of the network. Defines two concepts of maximum attack-bear capacity of the system and feasibility, establishes an extended Petri net attack model based on Attack Graph named AEPN to discuss network attack in two dimensions. Lastly, based on color Petri net and time Petri net, propose merger rules of the color Petri net without losing any information and simplify the system model scale at the same time. Propose inference rules of the time Petri net and finally propose a fuzzy color Petri net attack model based on time Petri net and its inference rules, which can describe the process of multi-attackers attacking the same target at the same time.In this paper, three kinds of models and corresponding reasoning algorithms are presented and verified with examples. Express the importance of researching attack model. In summary, the paper uses Attack Graph and Petri net to analyze the process of network attack, and open up new ways to study network attack models.