RAPID: Improving Capabilities to Measure the Robustness of Critical Communications Infrastructure: A Case Study of COVID-19 Quarantine-Induced Internet Performance

在新型冠状病毒疫情的危机下，互联网正面临前所未有的流量激增，这是由于使用基于云的远程办公和远程教育工具进行实时视频会议和在线课程所导致的。然而，由于测试服务器的位置，现有的网络吞吐量测量（速度测试）平台通常不能代表基于云的应用的性能。衡量云平台和互联网服务提供商之间的互连性能（云-ISP链接）是理解基于云的视频会议应用的用户体验性能的关键缺失部分。该项目将设计和部署新颖的实验来测量云-ISP链接的网络拥塞，跟踪强制性就地保护命令期间性能下降的演变，随着美国大部分劳动力重返工作场所， 该科学测量项目将解决三个挑战：1）基于拓扑测量和分析，在住宅接入和中转ISP中战略选择测试服务器，2）速度测试算法和行为的科学表征，以及3）正确解释测量结果。这些实验将利用各种商业平台部署的速度测试服务器（例如，Ookla）作为从托管高带宽应用程序的云平台进行吞吐量测量的Vantage位置。该实验的实施将提供纵向数据，以分析云-ISP互连的性能下降。该研究将通过收集和共享数据，揭示基于云的远程工作应用程序的体验质量，从而研究美国关键互联网基础设施在疫情期间的性能和可靠性。将生成可视化数据，以提供网络性能的纵向视图，并在地理和拓扑上以细粒度识别性能降级。 由此产生的工具和方法将有助于指导未来的基础设施改进，并提高监控日益增长的关键通信生态系统的能力。该项目的详细信息和数据可在www.example.com上获得https://webspeedtest.caida.org.This奖项反映了NSF的法定使命，并被认为值得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来支持。

In the crisis of the CoVID-19 pandemic, the Internet is facing unprecedented surges of traffic induced by the use of cloud-based telecommuting and remote education tools for real-time video meetings and online classes. However, existing network throughput measurement (speed test) platforms are often not representative of the performance of cloud-based applications due to the location of test servers. Measuring the performance of interconnections between cloud platforms and Internet Service Providers (cloud-ISP links) is a critical missing piece to understanding user-experienced performance of cloud-based videoconferencing applications.This project will design and deploy novel experiments to measure network congestion of cloud-ISP links, tracking evolution of performance degradations during mandatory sheltering-in-place orders, and as large segments of the U.S. work force transition back to their workplaces. This scientific measurement project will tackle three challenges: 1) strategic selection of test servers within residential access and transit ISPs, based on topology measurements and analytics, 2) scientific characterization of the algorithms and behavior of speed tests, and 3) correct interpretation measurement results. The experiments will leverage speed test servers deployed by various commercial platforms (e.g., Ookla) as vantage points to conduct throughput measurements from cloud platforms hosting high-bandwidth applications. Operationalizing this experiment will provide longitudinal data to analyze performance degradation of cloud-ISP interconnections.This research will enable studying the performance and reliability of critical Internet infrastructures in the U.S. during this pandemic, by collecting and sharing data that sheds light on the quality of experience of cloud-based telework applications. Visualized data will be generated to provide longitudinal views of network performance, and identify performance degradations at fine granularity, geographically and topologically. The resulting tools and methods will help guide future infrastructure improvements, and advance the ability to monitor an increasing critical communications ecosystem.Detail information and the data of this project is available at https://webspeedtest.caida.org.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.