论文标题:网络抗毁性研究及其在证券交易网络中的应用
Research on Network Survivability and Its Application in Stock Trading Network
论文作者
论文导师 王能,论文学位 博士,论文专业 系统分析与集成
论文单位 华东师范大学,点击次数 114,论文页数 114页File Size6885K
2007-10-01论文网 http://www.lw23.com/lunwen_119131137/
network survivability;; measure of survivability;; survivable network design;; stock trading network;; network nodes
网络在信息化社会中扮演着越来越重要的角色，社会经济、日常生活对网络的依赖性越来越大。一旦网络暂时中断，将给社会造成许多负面影响。尤其对一些关键领域，如金融、证券行业，网络中断不但使其业务停顿，还可能给用户带来不可挽回的经济损失，甚至给行业和社会带来重大冲击。因此，对网络抗毁性进行深入研究并最终寻找出一套提高网络抗毁性的理论和可操作的方法已成为网络研究的一个重要课题。 本论文在现有研究成果的基础上，对网络抗毁性测度、抗毁性网络设计、网络抗毁性原理和基于网络抗毁性原理的证券通信网络建设以及如何提高网络节点的抗毁性等方面的内容进行了深入地研究。作者首次提出了生成树优化算法求解有连通度和跳数限制的抗毁性网络设计问题，并通过实例对理论研究成果进行了验证。随后将研究成果运用到实际工程中，取得了良好的效果。 首先，论文介绍了课题的研究背景，并给出了网络抗毁性的定义。然后作者对目前网络抗毁性研究的两个主要方向——抗毁性网络设计和网络容量优化的各自研究内容进行了归纳总结，并分别进行了详细介绍。 在此基础上，在引入了网络抗毁性的测度等内容后，作者对影响网络抗毁性的各项因素进行了深入分析。针对各种复杂网络结构，作者自行设计了平均抗毁性系数、加权抗毁性系数、多条备份路径抗毁性系数和组合节点备份抗毁性系数等一系列网络抗毁性测度指标，以便对各种结构网络的抗毁性进行定量地分析。 在抗毁性网络设计问题方面，如何以较小的代价建设符合抗毁性要求的网络一直是该领域研究的主要目标。针对需要同时满足连通度及跳数限制的抗毁性网络设计问题，本文作者自行设计了采用生成树优化算法来求解这类问题。该算法从理论上有效地解决了此类网络设计问题。生成树优化算法对大型网络的规划、设计或对已有网络进行扩容、优化均具有较大的实际应用价值和可操作性。如果进一步结合网络扩容中投资费用问题进行后续研究和开发，本算法将可在工程中具备良好的实际应用价值。 证券行业对我国国民经济的贡献日益增多，影响也日渐增强。证券行业由于其行业的特殊性，对证券通信网络有着很高的抗毁性要求。建设一个具有较高抗毁性的证券通信网络，对于证券市场的健康发展有着十分重要的意义。 作者以证券行业的网络结构为实例对网络抗毁性理论研究成果加以验证。本文通过对证券公司的网络结构特点的介绍和分析，结合证券公司交易模式的转变，使用生成树优化算法计算出不同模式下证券公司内部通信网络理想的网络拓扑结构，然后通过与证券公司的实际网络结构的比较，验证了生成树优化算法结果的合理性。 上海证券交易所(上证所)是中国证券市场的核心，对证券市场有举足轻重的作用，它的的通信网络是整个证券通信网络的核心。本文对上海证券交易所的通信网络特点进行归纳和总结，对现有的高抗毁性的通信网络结构进行分析。作者总结出的经验和成果，对于其他类似的网络核心建设具有重要的借鉴作用。同时，在上证所目前良好的网络结构基础上，作者提出了如何进一步完善核心网络的建议。 在关注整个网络拓扑结构的抗毁性的同时，不能忽视网络节点的抗毁性建设。作者以自己在攻读博士学位期间负责的证券技术大厦网络规划、设计项目作为工程实例进行了介绍。在整个证券技术大厦通信接入及网络通信系统的规划、设计中，作者将网络抗毁性研究成果与自身多年工作经验总结出的提高网络抗毁性的方法相结合，提出了一套提高网络中重要节点整体抗毁性的方法，从多个角度系统性地提高网络节点的网络抗毁性，使得该大楼的网络抗毁性达到国内领先水平。 最后，作者对全文的主要内容和研究成果进行了总结，并对该研究领域的未来发展趋势作了分析和展望。
The Internet and other communication networks play a more important role in our life and we become much more dependent on networks. The malfunction in network will cause a temporary hold up to the network and a short time inconveniences to the user. Especially in the communication networks for certain area such as financial and stock trading, the malfunction will not only break their major business but also cause irretrievable loss to the network user, even bring unfavorable influence to the society. For this reason, an intensive research to the network survivability and the theory and operational approach to improve network survivability are important topics in network technology. On the basis of existing research, this paper has made an intensive research to the network survivability measure, survivable network design problem, network survivability theory applied in stock trading communication networks and the approach to improve network survivability of important key nodes. In this paper, we present an original algorithm of designing survivable network, and achieve a good result in real application using the theory combined with practical experience. In the beginning, the research background of this topic is introduced, and the definition of network survivability is given. As two major research areas of network survivability, the research content of survivable network design and network capacity optimization are summarized and introduced in detail. On the basis of this, factors which affecting the network survivability are intensively analyzed. According to deferent kinds of complicated network structure, a series of measure index of network survivability are designed in the purpose of quantitatively analyzing survivability of deferent kinds of network structure. Constructing a minimum cost network satisfying certain survivability requirements is the main target of survivable network design (SND) problem.In order to solve the SND problem with connectivity and hop constraint, an original spanning-tree optimization (STO) algorithm is designed which is able to solve the STO problem efficiently. The algorithm is useful and operational not only in planning and designing of large scope network but also in expanding and optimization of existing network. Given further research and development with investment in network expanding, the algorithm is valuable for practical application. The research achievement of network survivability is successfully used in construction of stock trading communication network. With continuous development of stock market, the securities business has achieved more contribution to the national economy and stronger influence. With its specialty, the securities business demands higher survivability to communication network. Constructing a communication network with high survivability is of great importance to the development of stock market. In this paper, the network structures of securities companies are introduced and analyzed. With the changing of trading pattern in securities companies, the demand of inner communication networks under different pattern is analyzed and the topological structure of network is advised. As the kernel of the China stock market, Shanghai Stock Exchange (SSE) plays an important role in it. And the communication network of SSE is also the kernel of the stock trading communication networks. In this paper, this network was intensively analyzed from the angel of network survivability. We discuss the characteristic, analysis the topological structure, and give some advices to expand business usage on the network. When the topological structure of the whole network has reached a fairly survivability, the survivable construction of important key nodes in network should not be ignored. After the analysis with networks of securities companies and SSE, the planning of Stock Technology Building network which is designed by the author is introduced in detail as an engineering example. During the planning and designing of access network and communication system of the whole building, using the research achievement of network survivability combined with practical measures concluded from long time working experience, the author systematically improved the survivability of important key nodes from many angels and achieved a good result. In the end, the main content and research achievement of this paper are concluded. The tendency of the research area is analyzed and forecasted.
|
