Collaborative Research: NeTS-NBD: Cross-Layer Design and Management of Translucent Optical Networks

随着社会对信息依赖程度的提高，现有的信息技术（IT）基础设施将超出其能力范围。实际的纤维可以承受至少十年以上的负荷；瓶颈是网络节点，即电子交换机和路由器。尽管人们普遍认为，在国家和大陆网络规模上，网络内的光电（O-E）再生是不可避免的，但仍有限制这种O-E再生的趋势。光传输和交换提供了对协议、比特率、调制格式等透明的优势，从而为基础设施提供了一条更顺畅的发展路径，毫无疑问，基础设施将从最初部署的框架发展起来。透明度引入了一个设计问题，这个问题目前还没有得到很好的研究或理解，即物理层和更高层之间的强交互作用。这个问题需要跨层设计方法，这在无线领域是标准的，但在光学领域还未被探索。该提案旨在通过汇集两个具有非常互补的专业知识的pi来推进这一关键领域的知识。提出的努力是设计和分析光纤传输网络管理系统，用于使用更强大但有问题的全光交换机的网络。该方法是一个整体的方法，结合了光纤和交换机的退化，并优化了资源的动态分配。研究计划包括两个主要任务。首先是网络管理系统的优化，通过光电路的功率、路由和波长分配来解决。其次是考虑网络组件和参数对物理层性能的影响来设计物理网络。大型光交换机的真实数据可供pi使用，并将用于测试和验证性能预测和新算法。研究结果影响到我们社会的各个方面，这些方面依赖于对信息日益增长的需求。它鼓励开发需要大量带宽的应用程序。它使工业界能够购买和控制他们的网络链接，从而以更实惠的价格为他们提供更多的带宽。它还促进了网络和通信两个领域的思想交流。更广泛的影响：本研究提高了参与研究的研究生和本科生的教育水平。寻求来自不同背景的学生。pi有从广泛的人才库中挑选学生的历史，为整个团队提供多样化的研究环境。研究通过毕业论文项目整合到本科教育中。研究结果也将被纳入pi教授的光网络和通信的研究生课程。这将确保精通跨层设计问题的研究人员和工程师源源不断。

The increase in our society's dependence on information will push the existing information technology (IT) infrastructure beyond its capability. The actual fibers can handle the load for at least a decade more; the bottleneck is the network nodes, i.e., the electronic switches and routers. Even as there is a general consensus thatopto-electronic (O-E) regeneration within the network is unavoidable at national and continental network scales, there is a push to limit such O-E regeneration. Optical transport and switching provide the advantage of transparency to protocols, bit-rates, modulation formats, etc., thus enabling a smoother evolution path for an infrastructure that will undoubtedly grow from its initially deployed framework. Transparency introduces a design issue that is not well studied or understood at this point, namely, the strong interaction between the physical and higher layers. The problem calls for cross-layer design approaches, which are standard in the wireless community but are largely unexplored in the optical community. This proposal seeks to advance the knowledge in this critical area by bringing together two PIs with very complementary expertise.The proposed effort is to design and analyze fiber transport network management systems for networks that use the more powerful but problematic all-optical switches. The approach is a holistic one, incorporating the degradations in the fibers and switches, and optimizing the dynamic allocation of resources.The research plan includes two major tasks. The first is the optimization of the network management system, which is tackled through power, routing, and wavelength assignment for optical circuits. The second effort is to design the physical network considering the impact the network components and parameters have on the physical layer performance. Real data on large optical switches are available to the PIs and will be used to test and validate the performance predictions and new algorithms.The results of the research affect every aspect of our society that depends on the ever-increasing need for information. It encourages the development of applications that require significant bandwidth. It empowers industry to purchase and control their networking links, thus giving them more bandwidth at a more affordable price. It also stimulates cross-fertilization of ideas from the two fields of networking and communications.Broader Impact: The research enhances the education of the graduate and undergraduate students participating in it. Students from diverse backgrounds are sought. The PIs have a history of selecting students from a broad pool to provide a diverse research environment for the whole group. Research is integrated into undergraduate education via senior thesis projects. The research results will also be incorporated in graduate classes that the PIs teach on optical networking and communications. This will ensure a steady stream of researchers and engineers well versed in cross-layer design issues.