Collaborative Research: IMR:MM-1B: Privacy in Internet Measurements Applied To WAN and Telematics

PIMAWAT（应用于广域网和远程信息处理的互联网测量中的隐私）项目将展示提供尊重最终用户隐私的数据网络数据集的新方法，同时仍然能够支持网络协议、安全、隐私和机器学习方面的新研究。主要观点是，今天通过广域网（WAN）发送的大多数数据都是加密的；因此，挑战在于演示哪些数据是加密的，检测并清除任何剩余的泄漏，最后在共享数据之前匿名化元数据（谁与谁交谈）。PIMAWAT的智力优势在于开发大规模匿名化网络流量的新方法，然后使用这些新算法评估潜在的数据泄漏，并证明可以在尊重隐私的同时清除真实世界的数据源以进行共享。PIMAWAT计划将研究人员先前的工作重点放在广域网数据流量上。如有可能，它还将探索汽车远程信息处理作为一个最近开发的数据集，它带来了独特的隐私机会和挑战，以设备（而不是人）为重点，但有地理位置和应用程序细节。PIMAWAT的更广泛影响将是通过新工具和尊重隐私的数据清洗的最佳实践，使收集和共享网络数据的潜力民主化。该项目的数据将为协议设计和网络安全提供计算机科学的新方法，应用人工智能和机器学习，并将在快速发展的车辆远程信息处理领域提供早期结果。PIMAWAT将提供新的工具、数据和实践，并鼓励其他研究人员、课堂和行业使用这些方法。PIMAWAT项目的网站将是https://ant.isi.edu/pimawat，其工具和数据集将在项目期间和完成后通过https://ant.isi.edu/datasets/提供。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 PIMAWAT (Privacy in Internet Measurements Applied to WAN And Telematics) project will demonstrate new methods to provide data networking datasets that respect end-user privacy, while still being able to support new research in network protocols, security, privacy, and machine learning. The main insight is that *most data today sent over the wide-area network (WAN) is encrypted*; thus, the challenge is to demonstrate what data is encrypted, detect and scrub any remaining leaks, and finally anonymize the metadata (who talks to whom) before sharing data.The intellectual merit of PIMAWAT will be to develop new methods to anonymize network traffic at scale, then use those new algorithms to evaluate potential data leakage, and demonstrate that real-world data sources can be scrubbed for sharing while respecting privacy. PIMAWAT plans to focus the investigator’s prior work on wide-area network data traffic. As possible, it will also explore vehicle telematics as a recently developing dataset that poses unique privacy opportunities and challenges, with a device (not person) focus, yet with geolocation and application details.The broader impacts of PIMAWAT will be to democratize the potential to collect and share network data through new tools and best-practices for privacy-respecting data scrubbing. Data from this project will enable new approaches in computer science for protocol design and cyber-security, applying AI and machine learning, and will provide early results in the rapidly evolving field of vehicle telematics. PIMAWAT will provide new tools, data, and practices, and encourage use of these methods by other researchers, in classrooms, and by industry.The PIMAWAT project website will be https://ant.isi.edu/pimawat and its tools and datasets will be provided through https://ant.isi.edu/datasets/ as they are developed during project and after it completes.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.