【作者】 俞艳苹；

【导师】 马建峰；

【作者基本信息】 西安电子科技大学 ， 计算机系统结构， 2005， 硕士

【摘要】 本文针对现有网络入侵防御手段本身存在难以解决的缺陷以及当前容忍入侵系统中自适应性和容忍故障／入侵能力较差的问题，提出了一种基于表决机制的容忍入侵系统模型。文中对支持该模型的核心机制即表决技术进行了深入的研究，设计了一种满足容忍入侵系统需要且行之有效的表决方法—自适应大数表决算法，使得系统即使在遭受到恶意攻击的情况下仍能提供安全、可靠的服务，并能有效地识别、屏蔽、隔离各种故障／入侵，从而达到了容忍入侵的目的。本系统将表决机制与重配置策略进行了有效地结合，使它具有更强的自适应性和灵活性，增加了攻击者进行连续攻击的难度。同时，由于系统采用了效率和可靠性较高的主动复制以及群组通信技术，因此为容忍入侵系统提供了安全可靠的消息多点传送及复制管理功能。通过对系统进行性能仿真和分析，结果表明，相比其它的表决方法，自适应大数表决算法安全性和可靠性更高。同时，整个系统具有很强的故障／入侵能力和自适应性，还有性能好、易于实现等优点，特别适合应用于对系统安全性、可靠性、可生存性，以及系统效率要求较高的分布式环境中。更多还原

【Abstract】 In order to improve the limitation of current intrusion defense measures, this paper researches and designs an intrusion tolerant system with voting mechanism, with emphasis on providing the ability to tolerate fault/intrusion, adaptability, security and good reliability. An adaptive majority voting algorithm is deeply presented to help our system maintain secure and reliable services, detect and remove the faulty and malicious replicas, after faults/intrusions happen, thus gaining the ends of intrusion tolerance. The voting mechanism integrates with reconfiguration strategies to better adaptability and resilience in our system. In addition, active replication and group communication with high efficiency and reliability are used to ensure secure message multicast and replication management. The results from simulation and analysis indicate that not only does the adaptive majority voting maintain a higher level of security and reliability; it does so while outputting the results and removing malicious replicas more correctly compared with other voting algorithms. The whole system keeps high performance and efficiency, adapts to be taken into practice. Obviously, this system is of great theoretical and realistic significance to our further study and development of intrusion tolerant systems.更多还原