NeTS: JUNO2: Resilience in Next-Generation Intelligent Optical Networks

世界上的电信基础设施以光纤为主导，因为它具有巨大的带宽。为了支持处于商业和日常生活中心的互联网，在基础设施方面已经并将继续进行巨额投资。然而，随着带宽需求的增长，网络变得越来越复杂，管理和防范自然灾害和其他灾害的难度也越来越大。因此，网络设计者越来越多地转向由基本功能的简单构建块组成的光节点。该项目旨在为网络基础设施提供抗故障的弹性，包括小规模（由于组件或系统退化）和大规模（例如，由于灾难）。将开发从组件级到网络级再到服务级的多个级别的弹性方案。这是一个与日本名古屋大学和香川大学的研究人员合作的项目。该合作将利用广泛的实验室设备进行原型设计和测试开发的技术，并与项目的日本合作者进行密切的行业互动。该项目的结果将提高光网络的可信度，而光网络是我们信息和通信基础设施的支柱。该研究解决了设计从组件级到网络级和服务级的多粒度可信光网络的挑战性问题。通过利用机器学习和马尔可夫决策过程等多种技术，将开发具有成本效益的架构和实用的智能算法，以应对多次故障，并提供服务级可用性保证。我们的研究目标是：(a)通过以经济有效的方式引入冗余，设计由不可靠组件组成的高可靠光节点。(b)设计智能算法，使其在网络层面的多重故障中存活。(c)设计可信赖的服务资源管理算法。(d)研制小型节点原型，并在系统试验台评价其性能。这将利用该项目的日本合作机构的实验室设施来实现。该项目将使新的理论和算法被纳入高级研究生课程，并为本科生开发新的实验项目。该项目将招收未被充分代表的少数民族学生，并包括广泛的K-12外展。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The world's telecommunication infrastructure is dominated by fiber optics because of its tremendous bandwidth. Huge investments have been and continue to be made in that infrastructure in order to support the Internet, which is at the center of business and daily life. As demand for bandwidth grows, however, the network becomes more complex and harder to manage and protect against natural and other disasters. Network designers are therefore increasingly turning toward optical nodes composed of simple building blocks of essential functions. This project aims to equip the network infrastructure with resilience against failures, both small-scale (due to component or system degradation) and large-scale (due to disasters, for example). Resilience schemes at multiple levels, ranging from component-level to network-level to service-level, will be developed. This is a collaborative project with researchers from Nagoya University and Kagawa University in Japan. The collaboration will leverage the extensive laboratory equipment for prototyping and testing the developed technology and the close interaction with industry that the project's Japanese collaborators have. The results of the project will increase the trustworthiness of optical networks, which form the backbone of our information and communications infrastructure. The research addresses challenging problems in designing trustworthy optical networks at multiple granularities ranging from component level to network level and to service level. Cost-efficient architectures and practical and intelligent algorithms for surviving multiple failures and for providing service-level availability guarantees will be developed by leveraging multiple techniques such as machine learning and Markov decision processes. Our research objectives are to:(a) Design highly reliable optical nodes that consist of unreliable components by introducing redundancy in a cost-effective manner.(b) Design intelligent algorithms for surviving multiple failures at the network level.(c) Design trustworthy service resource management algorithms.(d) Develop a small-scale node prototype and evaluate its performance in a system testbed. This will be achieved using the laboratory facilities at the project's Japanese partner institutions. This project will enable new theory and algorithms to be incorporated into advanced graduate classes and new experimental projects to be developed for undergraduate students. The project will recruit students from underrepresented minorities, and includes extensive K-12 outreach.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.