首页> 外文学位 >New results in the design and analysis of nonblocking switching networks.
【24h】

New results in the design and analysis of nonblocking switching networks.

机译:无阻塞交换网络设计和分析的新结果。

获取原文
获取原文并翻译 | 示例

摘要

Switching networks are the core component of modern switching devices for voice and data communications. The evolution of switching networks dates back to the early days of telephony. Advances in communication technologies, such as broadband communications and optical transmissions, as well as the explosion of demand on switching capacity have made switching network design an ever-more challenging task. Switching speed is the bottleneck at the core of the Internet and communication-intensive data centers. Consequently, characterizing the complexity of switching networks and constructing cost-effective switching networks under modern requirements are important problems both theoretically and practically.;In switching network theory, the first important class of questions needed to be addressed is of the type: "how complex is it to design a switching network satisfying such and such requirements?" In order to answer this type of questions, we first need to formulate a complexity model for the switching networks in the desired switching environment. This class of questions has been studied extensively in the classical environment of telephone (or circuit) switching for the past 60 years. Results and techniques obtained for circuit switching have proved to be very useful, with deep connections to other areas of Mathematics and Theoretical Computer Science. Under the newer switching environments such as broadband switching and optical switching, much less is known, however. This dissertation partly fills the gap by addressing two important complexity problems in broadband switching and optical switching. Briefly, we resolve a 20-year old open question regarding the optimal complexity of a multirate distributor, which is the complexity model representing rearrangable multicast switching network under the broadband environment. We also provide several complexity models for optical switching networks and show how to construct optical switching networks of optimal sizes.;The second class of questions in switching network studies is to design practical switching networks using commodity switching elements. This type of questions certainly has also been studied extensively in circuit switching. Many good constructions are known, such as the Clos network, Cantor network, Benes network, and so on, which have been widely used in practice. This dissertation continues the tradition by introducing combinatorial optimization techniques for analyzing the class of multi-log networks. Our maxflow-mincut and linear programming based techniques prove to be simpler and more effective than known ones in analyzing multi-log networks. New constraints such as crosstalk and fanout in optical switching can be incorporated into the combinatorial problem formulation and handled effectively by our techniques.
机译:交换网络是用于语音和数据通信的现代交换设备的核心组件。交换网络的发展可以追溯到电话的早期。通信技术的进步,例如宽带通信和光传输,以及对交换容量的需求激增,使得交换网络设计成为一项越来越具有挑战性的任务。交换速度是Internet和通信密集型数据中心的核心瓶颈。因此,表征交换网络的复杂性并在现代要求下构建具有成本效益的交换网络在理论上和实践上都是重要的问题。在交换网络理论中,需要解决的第一类重要问题是:设计满足这种要求的交换网络吗?”为了回答这类问题,我们首先需要为所需交换环境中的交换网络制定复杂度模型。在过去的60年中,在电话(或电路)交换的经典环境中对此类问题进行了广泛的研究。事实证明,与电路交换有关的结果和技术与数学和理论计算机科学的其他领域有着深厚的联系。然而,在诸如宽带交换和光交换之类的较新的交换环境下,人们所知甚少。本论文通过解决宽带交换和光交换中两个重要的复杂性问题,部分填补了空白。简而言之,我们解决了一个关于多速率分配器的最佳复杂度的20年未解决的问题,该问题是代表宽带环境下可重排的多播交换网络的复杂度模型。我们还为光交换网络提供了几种复杂性模型,并展示了如何构建最佳尺寸的光交换网络。交换网络研究中的第二类问题是使用商品交换元件设计实际的交换网络。当然,在电路切换中也已经广泛研究了这类问题。已知许多良好的构造,例如Clos网络,Cantor网络,Benes网络等,它们在实践中已被广泛使用。本文通过引入组合优化技术来分析多对数网络的类别,延续了传统。我们的基于maxflow-mincut和线性编程的技术在分析多对数网络方面比已知的技术更简单,更有效。诸如光交换中的串扰和扇出之类的新约束可以并入组合问题公式中,并通过我们的技术有效地处理。

著录项

  • 作者

    Wang, Yang.;

  • 作者单位

    State University of New York at Buffalo.;

  • 授予单位 State University of New York at Buffalo.;
  • 学科 Computer Science.
  • 学位 Ph.D.
  • 年度 2010
  • 页码 116 p.
  • 总页数 116
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号