IMR: MM-1C: An Extensible Platform for Asking Research Questions of High-Speed Network Links

在过去的十年中，网络速度急剧增加，超过了研究人员和运营商用于理解和保护网络的工具。除了流量之外，互联网流量本身也发生了根本性的变化——今天的大多数流量都是加密的，依赖于云基础设施或内容分发网络。对Internet流量的可见性对于理解性能特征、排除操作问题、执行Internet度量研究和回答基本安全问题至关重要。已经提出了许多用于恢复网络可见性的工具，但由于能够提出的问题类型的基本限制，很少有工具被广泛采用。重新表述后，提出的工具速度很快，但无法适应研究人员和操作人员在实践中需要提出的复杂问题。该项目建立了一个开源的网络测量平台，专门设计用于询问复杂的、现实世界的网络问题，由广泛的研究人员和ISP运营商定义。该平台运行在通用（x86）服务器上，并允许用户订阅分析所需的已解析网络流量子集。与之前的工作相反，该平台不受自定义查询语言的限制，而是建立在Rust编程语言提供的灵活性、性能和安全性之上。该项目的成功将使研究人员能够更轻松地进行互联网测量和安全研究，使运营商能够更好地了解和保护自己的网络。为了确保平台的长期可用性，研究团队将额外与外部研究小组合作，以了解生产网络中存在的各种操作限制，并在这些环境中部署测量平台。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。

Over the past decade, network speeds have increased dramatically, outpacing the tools that both researchers and operators used for understanding and securing networks. Beyond volume, Internet traffic itself has also fundamentally changed — most traffic today is encrypted and relies on cloud infrastructure or content distribution networks. Visibility into Internet traffic is critical for understanding performance characteristics, troubleshooting operational issues, performing Internet Measurement Research, and answering fundamental security questions. Many tools have been proposed for regaining network visibility, but few have seen widespread adoptions due to fundamental limitations in the types of questions they were capable of asking. Rephrased, proposed tools were fast, but could not accommodate the complex questions that researchers and operators need to ask in practice. This project builds an open source network measurement platform that is specifically designed to allow asking complex, real-world questions of networks, as defined by a broad set of researchers and ISP operators. The platform runs on general purpose (x86) servers and allows users to subscribe to subsets of parsed network traffic needed for their analysis. In contrast to prior work, the platform is not constrained to a custom query language, but instead builds on the flexibility, performance, and security provided by the Rust programming language. The project’s success will enable researchers to more easily pursue Internet Measurement and Security research as enable operators to better understand and secure their own networks. To ensure the platform’s long term usability, the research team will additionally partner with external research groups to understand the diverse set of operational constraints present in production networks and to deploy the measurement platform in those environments.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.