IMR:MT: Internet Routing Experiments for the Cloud Era

新兴应用将越来越强调互联网的延迟、可靠性、吞吐量和抖动，从视频会议到AR/VR(增强现实/虚拟现实)，再到自动驾驶车辆和交通控制，再到远程健康监测和医疗保健。这些应用程序越来越集中在少数大型云和内容提供商身上。他们在世界各地部署服务器，为短距离的客户提供服务；他们直接与许多边缘网络对等，直接提供服务。对于这一重要背景，学术研究人员缺乏现实的实验环境。现有试验台缺乏大型现代提供商的全球覆盖范围和丰富的互联能力，无法支持新兴应用程序所需的低延迟和高吞吐量流量。该项目将建立一个平台，使研究人员能够在忠于真实云提供商的环境中进行互联网路由实验。为了将这一重要背景下的研究民主化，本项目将在现有的对等试验床上构建一台新的软件路由器，支持高性能；该路由器将部署在六大洲的30个云站点，拥有近万个BGP对等点，大幅扩大对等网络的足迹；该项目将为研究人员构建和提供工具，以在这种环境下进行实验，控制云站点和对等设备之间交换的路由和流量，并测量和归档结果。该项目的主要影响将是在全球部署、高度互联、具有精细控制的云网络环境下实现对研究的民主化访问，这对当今的互联网和社会至关重要。通过向没有特权接触超级巨头和云提供商的研究人员开放这一研究领域，该项目将使能够塑造未来互联网和互联网应用的尖端研究成为可能，造福社会。我们构建的高性能软件路由器在许多其他环境中也将有用。该项目还将开发涉及这一关键领域的真实世界实验的课堂项目，帮助培训这一关键领域的学生。该项目将扩展对等试验台，并将扩展记录在对等网站上：https://peering.ee.columbia.edu/This奖反映了国家科学基金会的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Emerging applications will stress the Internet’s latency, reliability, throughput, and jitter more and more, from video conferencing, to AR/VR (augmented reality/virtual reality), to self-driving vehicles and traffic control, to remote health monitoring and health care. These applications are increasingly concentrated on a small number of large cloud and content providers. They deploy servers around the world to serve clients over short distances; and they peer directly with many edge networks to directly deliver services. Academic researchers lack realistic experimentation environments for this important setting. Existing testbeds lack the global footprint and rich interconnectivity of large modern providers, and they cannot support the low-latency and high-throughput traffic that emerging applications demand. This project will build a platform that enables researchers to conduct Internet routing experiments in an environment that is faithful to real-world cloud providers. To democratize research in this important setting, this project will build a new software router for existing PEERING testbed capable of supporting high performance; the routers will be deployed at 30 cloud sites across six continents, with nearly 10,000 BGP peers, dramatically expanding PEERING's footprint; and the project will build and make available tools for researchers to conduct experiments in this setting, controlling the routes and traffic exchanged between the cloud sites and the peers, and measuring and archiving the results.The primary impact of this project will be in democratizing access to research in the setting of a globally deployed, highly interconnected cloud network with granular control, one which is critical and central to today's Internet and hence society. By opening this research area to researchers without privileged access to hypergiants and cloud providers, the project will enable cutting-edge research that can shape the Internet and Internet applications of the future, benefiting society. The high-performance software router we build will also be useful in many other settings. The project will also develop class projects that involve real-world experiments in this setting, helping to train students in this critical area.The project will extend the PEERING testbed, and the extensions will be documented on the PEERING website: https://peering.ee.columbia.edu/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.