【作者】 胡广涛；

【导师】 党小超；

【作者基本信息】 西北师范大学 ， 计算机应用技术， 2011， 硕士

【摘要】 随着Internet技术的发展和日益普及,网络规模不断扩大,复杂性不断增加,异构程度也越来越高,使得现代网络管理系统通常要面对一种复杂的、大规模的、大量的信息数据以及多元化的网络设备。在这种网络环境下,传统的基于Client/Server的集中式网络管理系统己经不能完全满足网络管理的要求,需要引入新的管理方法和体系结构。移动Agent技术作为一种新型的分布式计算技术,在电子商务、网络管理、基于Web的信息收集与发布以及并行计算等方面都得到了广泛的应用,尤其是基于移动Agent的网络管理技术一直是该领域研究的热点。论文主要研究内容如下:(1)描述移动Agent相关技术,分析各主流移动Agent平台,分析其适用于构建新型分布式网络管理系统的原因和优势;设计一种基于移动Agent的网络管理模型,分析其工作流程以及对比于传统网管模型的优势所在;(2)移动Agent定位技术。将整个系统划分为多个域,使用网络管理站和域合作的方式实现移动Agent的定位服务。网络管理站作为整个系统的定位中心,每个移动Agent在建立和跨域迁移时都向它注册,进入域内的第一个节点作为该域的定位服务器,代理在域内进行迁移时,则只需记录代理节点对应的位置信息。实验证明这种方法能够取得较好的效果;(3)移动Agent旅行路径。在基本蚁群算法的基础上,本文给出了一种遗传-蚁群混合算法(Combination of Genetic and Ant Colony Algorithm:CGACA),它利用遗传算法的交叉与变异的优点,通过引入交叉算子和变异算子增强蚁群算法寻找全局最优解的能力和加快算法的收敛速度。实验证明,CGACA相对于遗传算法和蚁群算法,有更好的效果;(4)介绍了目前运行移动Agent的主机所面临的主要威胁和产生这些威胁的原因,同时也分析了Agent自身所面临的安全问题。更多还原

【Abstract】 As the development and popularization of Internet Technology, network scale is extending, network complexity is increasing and heterogeneity is enhanced .Under these circumstances, diversified network equipments should be adopted for the cunrrent network management system to process massive complicated data. The traditional centralized network management system based on C/S model couldn’t meet the requirements; therefore the new network management and system architecture are developed. As a new distributed computing technology, Mobile Agent Technology is widely applied in E-business, network management, information collection & release based on web and parallel computing; especially network management based on Mobile Agent Technology has been researched by many people. Major contents in this paper are listed as following:(1) Mobile Agent related techniques, the current platform for Mobile Agent and the advantages of Mobile Agent in constructing the new distributed network management system, A network management model based on Mobile Agent; the work procedure and compare with the traditional model.(2) Location mechanism in Mobile Agent. The system is divided into different domins, network management station and domain cooperating is uesed to implement the Mobile Agent location service.The network management station is regarded as the location centre among network management system and all the Mobile Agents have to register to that when they are established or interdomin migrating. Then the Mobile Agent selects the first node as the agency’s positioning server according to Mobile Agent’s migratory path, and the node in this domian only need to record the corresponding position information. Experiments proved that the improved method could obtain a better effect.(3) Mobile Agent traveling routes. The Combination of Genetic and Ant Colony Algorithm (CGACA) based on ant colony algotithm is designed. The algorithm possesses the advantages of crossover and mutation in genetic algorithm. It can enhance the ability to search the most optimal solution in ant colony algorithm and accelerate convergence by incorporating crossover operator and mutation operator. Experiments testified that CGACA is more effective than the genetic algorithm and the ant colony algorithm.(4) The major threats Mobile Agent hosts face with at present and the causes of these threats and the security problems in Moble Agent.更多还原