CAREER: Decentralizing Trust in Open Distributed Systems

允许任何用户将其节点注册为参与者的分布式系统必须解决两个主要挑战来实现可靠性：选择可信的参与者，以及检测和遏制异常节点的损害。本研究通过(A)提出一种新的系统体系结构F2F来解决这些挑战，该体系结构在确定可信任节点时结合了用户之间的分散信任关系，以及(B)设计了新的技术来在使用该新体系结构的具体应用环境中检测和遏制行为不良节点的缺陷。本研究通过两个案例研究系统的设计和实现证明了F2F的有效性：协同备份和审查规避。Friendstore通过将每个用户的数据存储在受信任邻居的子集上，提供了廉价而可靠的备份。这样的架构允许Friendstore使用成本较低的技术手段来检测故障并从故障中恢复，但会使充分利用存储资源变得更加困难。该项目使用编码技术解决了这一挑战，该编码技术允许节点同时高效地为多个邻居提供数据冗余。第二个案例研究系统，万花筒，通过代理网络转发流量，帮助受审查域内的用户访问被屏蔽的网站。万花筒使用一种新的协议在F2F信任图上传播代理身份，使得每个节点只能收集关于其他节点身份的有限信息。因此，少数恶意节点无法发现相当一部分的代理。该项目将生产的系统，如果成功部署，将为家庭用户提供廉价的在线存储，并帮助绕过数以百万计的受影响用户的审查。

Distributed systems that allow any user to enroll his node as a participantmust resolve two main challenges to achieve reliability: selecting trustworthyparticipants, and detecting and containing the damage of deviant nodes. Thisresearch addresses these challenges by (a) presenting a new systemarchitecture, F2F, that incorporates decentralized trust relationships amongusers in determining trustworthy nodes and (b) devising new techniques todetect and contain faults of misbehaving nodes in the context of concreteapplications using this new architecture. This research demonstrates the usefulness of F2F through the design andimplementation of two case-study systems: cooperative backup and censorshipcircumvention. Friendstore provides inexpensive and reliable backup by storingeach user's data on a subset of trusted neighbors. Such an architecture allowsFriendstore to use less expensive technical means to detect and recover fromfaults but makes it harder to fully utilize storage resources. This project addresses this challenge using coding techniques that allow a node toefficiently provide data redundancy for multiple neighbors simultaneously. Thesecond case study system, Kaleidoscope, helps users inside censored domainsaccess blocked websites by relaying traffic via a network of proxies.Kaleidoscope uses a novel protocol to disseminate proxy identities over the F2Ftrust graph so that each node can only collect limited information aboutothers' identities. A few malicious nodes, then, cannot discover a significantfraction of proxies. This project will produce systems that, if successfully deployed, provideinexpensive online storage to home users and help circumvent censorship formillions of affected users.