网络安全评估理论及其关键技术研究

发布时间：2018-07-26 13:38

【摘要】：在互联网迅速发展的同时,各种新型的网络攻击手段也在不断涌现,导致网络信息安全问题变得十分突出。因此,面对各种网络威胁,必须采取有效措施来保证网络系统的正常运行。但是,传统的被动型安全防御技术已无法满足人们的需要,于是国内外学者纷纷致力于研究主动的网络安全分析与评估方法,其目的是主动分析网络中存在的安全隐患,并根据分析结果采取适当措施来降低网络的安全风险,而如何准确高效地对网络安全性进行评估就显得尤为重要,已成为当前网络安全领域的一个研究热点。本文以计算机网络安全性为核心,采用Petri网、博弈论、随机过程、模糊数学等理论,对网络安全评估中涉及的评估模型构建以及安全评估方法等关键技术进行了深入的研究。本文的主要工作包括以下四个方面:首先,研究了网络攻击行为的建模技术。针对现有模型大多欠缺对并发性和协作性攻击过程描述能力的问题,提出一种基于广义随机着色Petri网(Generalized Stochastic Colored Petri Net,GSCPN)的网络安全评估模型(GSCPN based Network Security Assessment Model,GSCPN-NSAM)。该模型适于描述并发性和协作性攻击,可利用着色Petri网的颜色集来表示攻击相关属性,同时可以基于随机Petri网对系统的性能进行评估。给出了评估模型的相关性质、建立算法以及模型的正确性验证方法,并且对模型的复杂度进行了度量。考虑到实际环境中网络规模过大容易导致模型节点数过多的问题,在模型的构建过程中引入了层次化思想,通过性能等价化简来降低模型的复杂度。其次,研究了网络脆弱性的分析方法。针对传统网络脆弱性分析方法的不足,重点研究了基于GSCPN-NSAM模型的网络脆弱性分析方法。首先,提出一种基于GSCPN-NSAM模型的最佳攻击路径分析方法,该方法通过计算每条攻击路径的时间代价来预测最佳攻击路径,能够避免现有方法中计算攻击成功概率容易出现的问题,其分析结果可以指导网络管理人员在安全风险最大的路径上加强防御;其次,提出一种基于GSCPN-NSAM模型的网络安全加固措施制定方法,该方法引入主机节点利用率指数和主机节点关键度等概念,通过计算主机节点的关键度对网络中需要修补的脆弱节点进行排序,在此基础上根据最大节点关键度优先的原则逐步对目标网络进行安全加固,从而提高了网络整体的安全性。再次,研究了网络安全风险的评估方法。指出现有风险评估方法中存在模糊因素不便统计,以及不适用于对经验知识进行建模与推理的问题。为解决这一问题,本文提出一种基于模糊Petri网的网络安全风险评估方法。建立了网络安全风险评估指标体系,并根据该指标体系构建模糊Petri网模型。给出了一种基于模糊Petri网的系统风险模糊推理算法,该算法运用矩阵运算进行推理,充分地利用了模糊Petri网的并行处理能力,同时结合层次分析法,定性与定量分析相结合地评估网络系统的安全风险。与传统的风险综合评估方法相比,本文提出的方法在评估过程中还加入了对风险因素事件可信度的分析,从而使评估结果更加准确和客观。最后,从攻防博弈的角度研究了网络安全最优防御策略的选取问题。针对网络攻防双方在攻防博弈分析中无法对双方的损益情况做出准确判断的问题,将三角模糊数的概念引入到博弈模型,提出一种基于三角模糊矩阵博弈的最优防御策略选取方法。给出了基于三角模糊矩阵的博弈算法,该算法通过求解三角模糊矩阵博弈的纳什均衡,可以帮助防御者预测可能的攻击行为,以及选取最优的防御策略。此外,利用重复博弈理论对攻防双方存在的长期对抗关系进行了分析。实例分析表明,引入三角模糊概念更加符合实际情况,提高了分析结果的准确性和有效性。
[Abstract]:With the rapid development of the Internet, a variety of new network attack means are constantly emerging, causing the problem of network information security to become very prominent. Therefore, in the face of various network threats, effective measures must be taken to ensure the normal operation of the network system. However, the traditional passive security defense technology has been unable to meet the needs of people. As a result, scholars at home and abroad have devoted themselves to the study of active network security analysis and evaluation methods. The purpose is to analyze the security risks existing in the network actively, and to take appropriate measures to reduce the security risk of the network according to the results of the analysis, and how to evaluate the security of the network accurately and efficiently is particularly important. At the core of the network security, this paper takes the computer network security as the core, uses Petri net, game theory, random process, fuzzy mathematics and so on, and studies the key technologies such as the evaluation model construction and the security evaluation method involved in the network security assessment. The main work of this paper includes the following The four aspects: first, the modeling technology of network attack behavior is studied. Aiming at the problem that most of the existing models are lacking in the description ability of concurrency and cooperative attack process, a network security assessment model based on Generalized Stochastic Colored Petri Net (GSCPN) is proposed (GSCPN based Network Security). Assessment Model, GSCPN-NSAM). The model is suitable for describing concurrency and cooperative attacks. The color set of the colored Petri net can be used to represent the related attributes of the attack. At the same time, the performance of the system can be evaluated based on the random Petri net. The related properties of the evaluation model, the building algorithm and the correctness verification method of the model are given, and The complexity of the model is measured. Considering the problem that the network size is too large and the number of model nodes is too large in the actual environment, the hierarchical idea is introduced in the process of building the model, and the complexity of the model is reduced by the simplification of the performance equivalence. Secondly, the analysis method of network vulnerability is studied. The deficiency of the method of sex analysis is focused on the method of network vulnerability analysis based on GSCPN-NSAM model. First, an optimal attack path analysis method based on GSCPN-NSAM model is proposed. This method can predict the best attack path by calculating the time cost of each attack path, and can avoid the success of the calculation attack in the existing method. The analysis results can guide the network managers to strengthen the defense in the path of maximum security risk. Secondly, a method based on the GSCPN-NSAM model is proposed for the establishment of network security reinforcement measures. The method introduces the concept of the host node utilization index and the key degree of the main machine node, and calculates the host nodes. The key degree is to sort the vulnerable nodes which need to be repaired in the network. On this basis, the security of the target network is strengthened gradually according to the principle of maximum node key degree priority, thus improving the security of the network as a whole. Again, the evaluation method of the network security risk is studied. In order to solve this problem, this paper presents a network security risk assessment method based on fuzzy Petri net, establishes a network security risk assessment index system, and constructs a fuzzy Petri net model based on the index system. A fuzzy P based on fuzzy P is given. The fuzzy inference algorithm of system risk in ETRI network, which uses matrix operation to reasoning, makes full use of the parallel processing ability of fuzzy Petri net, and combines the analytic hierarchy process and qualitative and quantitative analysis to evaluate the security risk of the network system. Compared with the traditional comprehensive assessment method of wind risk, the method proposed in this paper is in the end. In the course of the evaluation, the reliability of the risk factors is also analyzed, so that the evaluation results are more accurate and objective. Finally, the selection of the optimal defense strategy for network security is studied from the angle of attack and defense game. The concept of triangular fuzzy number is introduced into the game model, and an optimal defense strategy selection method based on triangular fuzzy matrix game is proposed. A game algorithm based on triangular fuzzy matrix is given. By solving the Nash equilibrium of triangular fuzzy matrix game, the algorithm can help the defender to predict the possible attack behavior, and the selection of the algorithm. In addition, the repeated game theory is used to analyze the long-term confrontation relationship between the two parties in the attack and defense. The example analysis shows that the introduction of the concept of triangular fuzzy is more consistent with the actual situation, and improves the accuracy and effectiveness of the analysis results.
【学位授予单位】：解放军信息工程大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP393.08

【参考文献】