Next-Generation Optical Backbone Networks

今天我们的互联网很拥挤。 随着越来越多的用户使用互联网，以及他们对带宽密集型应用程序（Web浏览器、Java程序、视频会议等）的需求，增加， 迫切需要非常高带宽的传输设施。 设计更快的IP路由器和高速ATM交换机是重要的措施，但由于相应的网络结构仍然受到电子速度的限制，因此它们是“停止间隙”的努力。 为了将互联网提升到下一代，需要一场网络技术革命，将其现有能力（特别是带宽）提升2到3个数量级！ 幸运的是，光纤是我们的“救星”。考虑到单模光纤的潜在带宽接近50 Tbps，应该尽一切努力利用这一巨大的资源。 波分复用（WDM）可以通过要求每个终端用户的设备仅以电子速率操作来利用该带宽，但是来自不同终端用户的多个WDM信道在同一光纤上被复用。 该研究项目的目标是使用光WDM构建下一代网络的骨干。 具体而言，将讨论以下研究主题： * WDM网络控制和管理（NC M），以建立一个 虚拟"可重新配置"的互联网， 用户的需求，与相应的研究挑战， 开发备用路径路由，故障管理， 以及控制信道管理机制。 * 光树：多播光骨干网，可以有效地 不仅支持许多新兴的组播应用，而且 为某些单播应用程序提供更快的服务， 连接端点之间的“平均跳距”。 * IP over WDM，使下一代网络真正具有可扩展性， 也就是说，它可以与我们现有的互联网基础设施顺畅地工作。 * 打击物理层损伤，使许多可取的 在采用智能网络的同时， 算法来“纠正”它的一些不太理想的属性。

Our Internet is congested today. As more users get on the Internet, and as their demand for bandwidth-intensive applications (web browsers, java programs, video conferencing, etc.) increases, there emerges an acute need for very high-bandwidth transport facilities. The design of faster IP routers and high-speed ATM switches are important measures, but they are ``stop-gap'' efforts since the corresponding network architectures are still limited by electronic speed. To elevate the Internet to the next generation, a networking technology revolution is needed that can scale up its existing capability (particularly its bandwidth) by 2 to 3 orders of magnitude! Fortunately, fiber optics is our ``savior.'' Given that a single-mode fiber's potential bandwidth is nearly 50 Tbps, every effort should be made to tap into this huge resource. Wavelength-division multiplexing (WDM) can exploit this bandwidth by requiring that each end-user's equipment operate only at electronic rate, but multiple WDM channels from different end-users are multiplexed on the same fiber. This research project is targeted towards building the backbone of the next generation of our networks using optical WDM. Specifically, the following research topics will be addressed: * WDM Network Control and Management (NC&M), in order to build a virtually ``reconfigurable'' Internet that allocates bandwidth to users on demand, with the corresponding research challenges being development of alternate-path routing, fault management, and control channel management mechanisms. * Light-Trees: Multicast Optical Backbone, that can efficiently support not only many emerging multicast applications, but also provide faster service to some unicast applications by reducing the ``average hop distance'' between end points of connections. * IP over WDM, so that the next-generation network is truly scalable, i.e., it can work smoothly with our existing Internet infrastructure. * Combatting Physical Laye r Impairments, so that many of the desirable properties of optics can be exploited while employing smart network algorithms to ``correct" for some of its not-so-desirable properties.