Low Latency, Scalable Photonic Packet Switching Fabric

低延迟、可扩展的光子数据包交换结构

• 批准号: 0322813
• 负责人: Keren Bergman
• 金额: $ 27万
• 依托单位: Columbia University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2003
• 资助国家: 美国
• 起止时间: 2003-08-15 至 2007-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0322813&HistoricalAwards=false
• 关键词: Low; Latency; Scalable; Photonic; Packet

The goal of the proposed research program is to develop a photonic packet switching fabric capable of interconnecting large-scale systems (with 2000 ports) in a closely coupled fashion featuring latency that is dominated by time-of-flight. An integrative photonic interconnection network will be demonstrated. The underlying photonic switching fabric is based on a novel architecture termed the Data Vortex, that utilizes hierarchal topology, simplified traffic control, and synchronous timing to reduce the necessary routing logic operations and buffering. These features drastically simplify the optical packet routing and facilitate the implementation the Data Vortex with commercially available photonic components. The optical packets are encoded in a wavelength division multiplexed (WDM) fashion to provide the optical data rate interfacing with the routing control and to achieve a minimum latency data exchange between physically separated memory and processor. Optical packets propagate through semiconductor optical amplifier (SOA) switch elements optimized for switching speed and minimal cross gain modulation. The integrated system will include the interfacing between the electronic processor/memory elements and the optical data streams in the switching fabric. Various traffic patterns will be tested and demonstrated including asynchronous and bursty packet distributions.The research program will provide unique graduate students training in a multifaceted complex systems program. The participating students will carry out the research as a team with each owning responsibility for the various subsystems or network levels. The close interdependence between the sub-projects will require each student to gain an understanding of all the system aspects. In this environment the students necessarily depend on each other to succeed fostering teamwork. Additionally, a new course on optical and computer networking for undergraduates is being introduced to the Computer/Electrical Engineering curriculum. This new course will be geared toward undergraduates with backgrounds in Electrical or Computer Engineering but not necessarily in optics. Interested undergraduate students particularly those from underrepresented groups will be drawn from the course to participate in this research program.

拟议的研究计划的目标是开发一种光子分组交换结构，能够以紧密耦合的方式互连大型系统（具有2000个端口），其延迟主要由飞行时间决定。一个集成的光子互连网络将被证明。底层的光子交换结构是基于一种新的架构，称为数据漩涡，利用分层拓扑结构，简化的流量控制和同步定时，以减少必要的路由逻辑操作和缓冲。这些功能大大简化了光分组路由，并促进了数据涡流与商业上可用的光子组件的实施。光分组以波分复用（WDM）方式编码，以提供与路由控制接口的光数据速率，并实现物理上分离的存储器和处理器之间的最小延迟数据交换。光分组通过针对切换速度和最小交叉增益调制而优化的半导体光放大器（SOA）切换元件传播。集成系统将包括电子处理器/存储器元件与交换结构中的光学数据流之间的接口。各种流量模式将被测试和演示，包括异步和突发数据包分布。该研究计划将提供独特的研究生培训，在一个多方面的复杂系统计划。参与的学生将作为一个团队进行研究，每个人都负责各个子系统或网络级别。子项目之间的密切相互依赖性将要求每个学生获得所有系统方面的理解。在这种环境中，学生必须相互依赖，以成功地培养团队合作。此外，在计算机/电气工程课程中，正在为本科生开设一门关于光学和计算机网络的新课程。这门新课程将面向具有电气或计算机工程背景的本科生，但不一定是光学背景。感兴趣的本科生，特别是那些来自代表性不足的群体将被从课程中吸引参加这项研究计划。