A Systematic Approach to the Design of Cost-Effective, High Performance Switching Architectures

设计经济高效、高性能交换架构的系统方法

• 批准号: 9502669
• 负责人: Ellen Zegura
• 金额: $ 15.16万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1995
• 资助国家: 美国
• 起止时间: 1995-06-15 至 1999-11-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9502669&HistoricalAwards=false
• 关键词: Systematic; Approach; Design; Cost; Effective

This proposal aims to answer a set of critical questions dealing with communication systems by way of the development and use of a simulation and implementation modeling environment for virtual circuit switches. The first question concerns systematic, standardized performance comparisons. Using relevant traffic models and trace data from actual applications, this study is standardizing performance measurements, eliminating the apples-to- oranges comparisons that result when considering performance results from different research groups. The second question concerns cost and implementation modeling. This work is integrating the design of system architecture and implementation. The particular focus is on electronic implementation and multichip module packaging technology. The third question concerns the role of architecture in providing quality of service guarantees. The usual abstraction of the switch as a queue hides too much detail about internal architecture that is relevant to traffic behavior when traversing a switch. This work is characterizing the effect on traffic of different switch architectures, and developing switch mechalisms to work in concert with other mechanisms in providing quality of service guarantees. The educational part of the project is using the simulation and implementation modeling environment as a key component of a graduate course in High Speed Switching Networks. In addition, portions of the research amp appear in two undergraduate courses, one on VLSI Design and the other on Modeling. Other education efforts include significant mentoring activity and ongoing improvement in teaching.

该建议的目的是回答一组关键问题处理通信系统的方式的开发和使用的虚拟电路交换机的仿真和实施建模环境。 第一个问题涉及系统的、标准化的业绩比较。 使用相关的流量模型和来自实际应用程序的跟踪数据，这项研究正在标准化性能测量，消除了考虑不同研究小组的性能结果时产生的苹果到橘子的比较。 第二个问题涉及成本和实现建模。 本工作是系统架构设计与实现的集成。 特别关注的是电子实现和多芯片模块封装技术。 第三个问题涉及体系结构在提供服务质量保证方面的作用。 通常将交换机抽象为队列，隐藏了太多关于内部架构的细节，这些细节与通过交换机时的流量行为有关。 这项工作的特点是不同的交换机体系结构的流量的影响，并开发交换机mechalisms与其他机制在提供服务质量保证。 该项目的教育部分是使用仿真和实现建模环境作为高速交换网络研究生课程的关键组成部分。 此外，部分的研究放大器出现在两个本科课程，一个关于超大规模集成电路设计和其他建模。 其他教育工作包括开展重要的辅导活动和不断改进教学。