Distributed Tree Infrastructure for Peer-to-Peer Systems

点对点网络是没有任何中央权威的分布式系统，用于有效管理和定位共享资源。与将负载集中在少数服务器机器上的传统客户机-服务器体系结构不同，点对点系统通过在参与者之间分配维护系统的成本来提供固有的可伸缩性。同时，点对点系统为高效和健壮的实现带来了新的挑战，因为它们是由不可信和不可靠的组件构建的，并且只能作为整个系统的紧急属性提供生存能力。该项目研究大规模分布式数据结构在点对点系统中的应用，重点是效率、可伸缩性、容错和恢复。该项目的主要目标包括(a)设计和实施在使用网络和使用本地存储之间进行最佳权衡的机制；(b)设计和实现快速构建大规模分布式数据结构的机制；(c)使用提供“范围查询”的数据结构，其中数据结构查找与所需对象最接近的匹配，而不仅仅是精确匹配，以支持版本控制，复制和近似匹配。

Peer-to-peer networks are distributed systems without any centralauthority that are used for efficient management and location ofshared resources. Unlike traditional client-server architectures,which concentrate load on a small number of server machines,peer-to-peer systems provide inherent scalability by distributing thecost of maintaining the system across its participants. At the sametime, peer-to-peer systems create new challenges for efficient androbust implementation, as they are built from untrusted and unreliablecomponents and can provide survivability only as an emergent propertyof the system as a whole.The project studies applications of large-scale distributed datastructures to peer-to-peer systems, with an emphasis on efficiency,scalability, fault-tolerance, and recovery. Major goals of theproject include (a) design and implementation of mechanisms foroptimal trade-offs between use of the network and use of localstorage; (b) design and implementation of mechanisms for very fastconstruction of large-scale distributed data structures; and (c) useof data structures providing ``range queries,'' in which the datastructure finds nearest matches to a desired object instead of justexact matches, to support versioning, replication, and approximatematching.