Abstract: It is necessary to build spatial index in order to managing complex spatial data in Peer-to-Peer（P2P）environments.This paper analyzes and summarizes the related studies，designs a new index model named P2PQR-tree which uses distributed Quad-tree and local R*-tree.P2PQR-tree applies Quad-tree techniques into P2P environments and uses replication strategy to get the load balance purpose.This paper gives the constructing algorithm of the index and gives the performance analysis of the algorithm.P2PQR-tree has some advantages to the old methods，for example，its data management is more reasonable，it can support metadata management better，implement rights control easier，reduce changes of distributed index，and adapt to dynamic character of P2P network better.

Key words: Peer-to-Peer, spatial data, Quad-trees, R*-trees

摘要： 在Peer-to-Peer（P2P）环境中管理复杂的空间数据，首先需要建立空间数据索引。对相关研究进行了分析总结，设计出了一种分布式四叉树与本地R*树相结合的新索引模型P2PQR-tree。P2PQR-tree将四叉树技术应用于P2P环境，并利用复制策略实现了负载均衡。给出了索引生成算法，并对算法性能进行了分析。P2PQR-tree数据组织更加合理、能够更好地支持元数据管理、易于实现权限控制、减少了分布式索引变化，更好地适应了对等网络的动态特性。

关键词: 对等计算, 空间数据, 四叉树, R*树