NeTS: Small: Broadband Service Reliability: Characterization and Improvement

本项目研究宽带业务的可靠性，探讨可靠性与用户体验质量的关系，研究基于边缘的可靠性提升方案。它开发技术和调查指标来研究和描述可靠性问题，以及算法来诊断和定位问题的来源，以现有的基础设施和作为联邦通信委员会“衡量宽带美国”工作的一部分收集的数据为基础。本研究探讨了可靠性问题对用户体验质量（QoE）的影响，探索了QoE可能的代理指标（如会话时间、流量需求）的价值，以及自然实验和相关设计的应用，以在控制大多数混杂因素的同时大规模地理解它们之间的关系。了解可靠性问题、它们的影响和最常见的原因应该有助于开发易于部署的解决方案，从而提高网络边缘的服务可靠性。随着住宅宽带网络的可用性和性能的提高，消费者继续将关键服务（如电话、电视和家庭监控）迁移到over-the-top替代方案。这种迁移伴随着用户对宽带可靠性的期望不断提高，这是消费者体验的关键因素。尽管可靠性工程有着丰富的历史，但研究人员对宽带环境下的可靠性的理解非常有限，从最合适的表征指标，到收集测量或评估其对用户体验质量影响的方法。当可靠性问题出现时，对它们进行故障排除几乎是不可能的——没有确定问题根源的原则性方法。其结果是增加了用户的挫败感，增加了Internet服务提供商的管理成本，并最终导致应用程序性能不佳。该项目加强了本科生和研究生的教学，让本科生参与研究项目，接触到行业，为研究人员提供工具和首创的数据集，并为联邦和州电信监管机构提供数据收集平台和数据。我们将开发的算法、工具、技术和实验方法将帮助用户、运营商和政策制定者为未来的投资提供更坚实的统计基础。

This project studies the reliability of broadband services, explores the relation between reliability and users' quality of experience and investigates edge-based solutions to improve reliability. It develops techniques and investigates metrics to study and characterize reliability problems, and algorithms to diagnose and localize their source, building on existing infrastructure and data collected as part of the Federal Communications Commission's Measuring Broadband America effort. This investigation explores the impact of reliability problems on users' Quality of Experience (QoE), exploring the value of possible proxy metrics for QoE (e.g., session time, traffic demand) and the application of natural experiments and related designs to understand their relationship at scale while controlling for most confounding factors. An understanding of reliability problems, their impact and most common causes should help in the development of readily deployable solutions that might improve service reliability at the network edge.As residential broadband network availability and performance improves, consumers continue to migrate key services such as telephony, television and home monitoring to over-the-top alternatives. This migration comes with rising users' expectations regarding broadband reliability, which is a key factor of consumer experience. Despite the rich history of reliability engineering, researcher have very limited understanding of reliability in the context of broadband, from the most appropriate metrics for characterizing it, to approaches for gathering measurements or assessing their impact on users' quality of experience. When reliability problems appear, troubleshooting them is nearly impossible -- there is no principled approach to determining the source of the problem. The outcome is increased frustration for users, growing management costs for Internet service providers, and, ultimately, underperforming applications. The project enhances undergraduate and graduate teaching, involves undergraduates in research projects, reaches out to industry, provides tools and first-of-its-kind datasets to researchers, and offers a data gathering platform and data for the benefit of Federal and State telecom regulators. The algorithms, tools, techniques and experimental approaches we will develop will help users, operators and policymakers providing a firmer statistical footing for future investment.