NeTS: Small: Cross Layer Control of Dynamic Optical Networks - from Theory to Experimentation

当前的光网络大多在最低光层静态地提供。 将通信基础设施迁移到动态配置的光基础设施将允许按需配置高带宽电路，以支持科学，商业，媒体和国家安全。 动态光网络将有助于弥合数字鸿沟，并支持带宽密集型的变革性应用，如用于教育和远程医疗的远程呈现的3D视频。 这种网络的物理层组件是从动态光学设备和技术的研究中产生的，并提供了独特的机会。然而，在这方面，由于共享同一光纤的多条光路的相互作用以及适应具有不同损伤的多种光纤类型和物理基础设施的挑战，在考虑物理层传输质量和更高层要求的同时控制动态光网络提出了重大挑战。用于广域和数据中心光网络的分层网络控制算法。该算法依赖于实时光性能监视器（OPM）的能力，并允许通过动态网络配置、再生以及功率、带宽和调制控制来有效地使用光资源。 由于广域网中动态操作的影响在分析上是棘手的，该项目开发了一个优化框架，而不是采用理论模型和分析解决方案，使用直接输入的OPM测量。对于数据中心网络，该项目正在设计网络配置方案，允许物理支持各种 *-cast（单播、任播、组播等）。更高层次的要求。该项目包括让不同的学生群体参与到哈莱姆社区的跨层次研究和外联活动中。

Current optical networks are mostly statically provisioned at the lowest optical layer. Moving the communications infrastructure to a dynamically provisioned optical infrastructure will allow on-demand provisioning of high-bandwidth circuits in support of science, commerce, media, and national security. Dynamic optical networks will contribute to bridging the digital divide and supporting bandwidth intensive transformative applications such as 3-D video for telepresence in education and telemedicine. The physical layer components of such a network are emerging from research on dynamic optical devices and technologies and offer unique opportunities. However, controlling dynamic optical networks while considering the physical layer Quality of Transmission and the higher layer requirements poses major challenges due to both the interactions of multiple light paths sharing the same fiber and the challenges of adapting to multiple fiber types and physical infrastructures with varying impairments.This project contributes to the deployment of dynamic optical networks by developing cross-layered network control algorithms for wide-area and data center optical networks. The algorithms rely on the capabilities of real-time optical performance monitors (OPMs) and allow efficient use of the optical resources through dynamic network configuration, regeneration, and power, bandwidth, and modulation control. Since the effects of dynamic operation in wide-area networks are analytically intractable, the project develops an optimization framework that, rather than employing theoretical models and analytical solutions, uses direct input from OPM measurements. For data center networks, the project is designing network configuration schemes that allow physical support of various *-cast (unicast, anycast, multicast, etc.) higher layer requirements. This project includes engaging diverse student populations in the cross-layered research and outreach activities to the Harlem community.