The Research On Key Evaluation Technique Of Network Attack Effectiveness

Effectiveness evaluation of network attacks is of great significance to the network attack and defense researches. Researches on the theories and methods of network attack effectiveness evaluation are helpful to make up the deficiencies of the network attack and defense researches and network security infrastructure in our country. By the qualitative and quantitative evaluation of the effectiveness of network attacks, on the one hand, it can help the attackers to test the effectiveness of attack actions and to promote the development of new attack techniques and methods; on the other hand, it can help the defenders to establish effective security policies of network and to promote their security. In this thesis, we carry out researches on three important aspects involved in network attack effectiveness evaluation process, including network attack effectiveness evaluation criteria, network attack modeling and evaluation methods. The main contributions and innovation points in the thesis could be summarized as follows:(1) A new criteria system of attack effectiveness evaluation from the perspective of network attack process is established. In this thesis, we classify attacks according to the network attack process and analyze each type of attacks. Then, we establish an criteria system composed by 84 criteria. It meets the requirement of phased evaluation and has the detectable and operable features so that the interruptions of the human factor are reduced and the objectivity of result is improved.(2) The Enhanced LAMBDA, an expanding attack description language, is put forward. Based on the LAMBDA language, this enhanced language expands its expressiveness capability on time constraints and effectiveness constraints. The results of the experiment indicate that Enhanced LAMBDA could accurately describe the complex combined attacks and effectively support the effectiveness evaluation of combined attacks.(3) An improved method for network attack effectiveness evaluation based on rough set theory is proposed. From the aspect of the data discretization, this thesis proposes an improved integrated discretization algorithm based on breakpoint importance. In respect of the weight determination, this thesis proposes an integrated weights determination method based on sum of squares of deviations.(4) An improved method for network attack effectiveness evaluation based on grey theory is put forth. It solves the problem that the evaluation model of network attack effectiveness, which is based on general grey clustering, can not accurately evaluate the object when there is no significant difference among the clustering coefficients.