摘要： 提出一种基于负载均衡的分层副本定位方法――RepliLoc。从副本信息存储和副本定位计算2种负载考虑，将副本定位问题划分为社区层和社区联合层。分别采用应用层广播方式和基于P2P资源定位Tapestry前缀匹配路由机制，实现副本局部和全局定位。通过哈希和前缀匹配方法将副本信息均衡存放，实现了存储负载均衡，通过社区将副本定位计算局部化，实现了计算负载均衡。

关键词: 数据网格, 副本定位, 负载均衡, 小世界, 前缀匹配

Abstract: This paper presents a hierarchical replica location approach based on load balance——RepliLoc. In RepliLoc, replica location is divided into two layers, one is community layer, and the other community alliance layer. Application layer broadcast is used to locate replica on community layer and matching prefix routing algorithms of Tapestry is used to achieve replica location on community alliance layer. Replica location messages are distributed evenly on each node by usiing Hash function and matching prefix to achieve storage load balance. Replica location is localized by community to achieve computational load balance.

Key words: data grid, replica location, load balance, small world, prefix matching

中图分类号: 

• TP393